blaisepascalmagazine 44 uk

PRINTED ISSUE PRICE 15.00

DOWNLOAD ISSUE PRICE 7 .50

BLAISE PASCAL MAGAZINEBLAISE PASCAL MAGAZINEBLAISE PASCAL MAGAZINE 444444

D E L P H I, L A Z A R U S, O X Y G E N E, S M A R T M O B I L E,

A N D P A S C A L R E L A T E D L A N G U A G E S

A N D R O I D, I O S, M A C , W I N D O W S & L I N U X

OVERVIEW OF DELPHI TO DELPHI HISTORY

EUCLIDES

AN AGE PUZZLE

BY DAVID DIRKSE

ARDUINO: THE VISUINO PROJECT - PART 4

INTERNET OF THINGS - WITH ARDUINO AND DELPHI

USE ETHERNET SHIELD, PROGRAM IT WITH VISUINO,

CONNECT FROM A DELPHI APPLICATION

OVER THE LOCAL NETWORK OR INTERNET

BY BOIAN MITOV

DATABASE WORKBENCH 5

THE SWISS ARMY KNIFE FOR DATABASES

BY PETER VAN DER SMAN

TIPS AND TRICKS WITH KBMMEMTABLE

BY KIM MADSEN

CONTENTS

BLAISE PASCAL MAGAZINEBLAISE PASCAL MAGAZINE

Articles

D E L P H I, L A Z A R U S, S M A R T M O B I L E S T U D I O,

A N D P A S C A L R E L A T E D L A N G U A G E S

F O R A N D R O I D, I O S, M A C, W I N D O W S & L I N U X

Issue Nr 6 2015 BLAISE PASCAL MAGAZINE

444444

OVERVIEW OF DELPHI TO DELPHI HISTORY PAGE 6

EUCLIDES PAGE 7

AN AGE PUZZLE PAGE 8

BY DAVID DIRKSE

ARDUINO: THE VISUINO PROJECT - PART 4

INTERNET OF THINGS - WITH ARDUINO AND DELPHI

USE ETHERNET SHIELD, PROGRAM IT WITH VISUINO,

CONNECT FROM A DELPHI APPLICATION

OVER THE LOCAL NETWORK OR INTERNET

BY BOIAN MITOV PAGE 12

DATABASE WORKBENCH 5

THE SWISS ARMY KNIFE FOR DATABASES

BY PETER VAN DER SMAN PAGE 28

TIPS AND TRICKS WITH KBMMEMTABLE

BY KIM MADSEN PAGE 39

Advertisers

BETTER OFFICE

COMPONENTS 4 DEVELOPERS 48

COMPUTER MATH & GAMES 4/5

DANIEL TETI DELPHI COOKBOOK 11

EVERS 36

NEW BLAISE LIBRARY 38

PASCON DELPHI 37

UPSCENE 35

VISUINO MITOV 27

Publisher: Foundation for Supporting the Pascal Programming Language in collaboration with the Dutch Pascal User Group (Pascal Gebruikers Groep)

Stichting Ondersteuning Programmeertaal Pascal

2

All material published in Blaise Pascal is copyright SOPP Stichting Ondersteuning Programeertaal Pascal unless otherwise noted and may

not be copied, distributed or republished without written permission. Authors agree that code associated with their articles will be made

available to subscribers after publication by placing it on the website of the PGG for download, and that articles and code will be placed on

distributable data storage media. Use of program listings by subscribers for research and study purposes is allowed, but not for commercial

purposes. Commercial use of program listings and code is prohibited without the written permission of the author.

Copyright notice

Editors

Peter Bijlsma, W. (Wim) van Ingen Schenau, Rik Smit,

Correctors

Howard Page-Clark, James D. Duff

Trademarks

All trademarks used are acknowledged as the property of their respective owners.

Caveat Whilst we endeavour to ensure that what is published in the magazine is correct, we cannot accept responsibility for any errors or omissions.

If you notice something which may be incorrect, please contact the Editor and we will publish a correction where relevant.

Subscriptions ( 2013 prices )

1: Printed version: subscription 65.-- Incl. VAT 6 % (including code, programs and printed magazine,

10 issues per year excluding postage).

2: Electronic - non printed subscription 45.-- Incl. VAT 21% (including code, programs and download magazine)

Subscriptions can be taken out online at www.blaisepascal.eu or by written order, or by sending an email to [email protected]

Subscriptions can start at any date. All issues published in the calendar year of the subscription will be sent as well.

Subscriptions run 365 days. Subscriptions will not be prolonged without notice. Receipt of payment will be sent by email.

Subscriptions can be paid by sending the payment to:

ABN AMRO Bank Account no. 44 19 60 863 or by credit card: Paypal

Name: Pro Pascal Foundation-Foundation for Supporting the Pascal Programming Language (Stichting Ondersteuning Programeertaal Pascal)

IBAN: NL82 ABNA 0441960863 BIC ABNANL2A VAT no.: 81 42 54 147 (Stichting Programmeertaal Pascal)

Subscription department Edelstenenbaan 21 / 3402 XA IJsselstein, The Netherlands / Tel.: + 31 (0) 30 890.66.44 / Mobile: + 31 (0) 6 21.23.62.68

[email protected]

Editor - in - chief

Detlef D. Overbeek, Netherlands Tel.: +31 (0)30 890.66.44 / Mobile: +31 (0)6 21.23.62.68

News and Press Releases email only to [email protected]

Authors - Christian name in alphabethical order

A Andrea Raimondi ,

B Peter Bijlsma, Dmitry BoyarintsevStephen Ball,

C Marco Cant, Michal Van Canneyt,

D David Dirkse, Daniele Teti

F Bruno Fierens

G Primo Gabrijeli, Mattias Gaertner

H Fikret Hasovic

J Cary Jensen

L Wagner R. Landgraf, Sergey Lyubeznyy

K Max Kleiner

M Kim Madsen, Felipe Monteiro de Cavalho

N Jeremy North,

O Inoussa Ouedraogo

P Howard Page-Clark,

S Rik Smit, Bob Swart,

Z Siegfried Zuhr

Stephen Ball

http://delphiaball.co.uk

@DelphiABall

Peter Bijlsma

-Editor peter @ blaisepascal.eu

Michal Van Canneyt,

michael @ freepascal.org

Marco Cant

www.marcocantu.com

marco.cantu @ gmail.com

David Dirkse

www.davdata.nl

E-mail: David @ davdata.nl

Bruno Fierens

www.tmssoftware.com

bruno.fierens @ tmssoftware.com

Primo Gabrijeli

www.

primoz @ gabrijelcic.org

Cary Jensen

www.jensendatasystems.com

http://caryjensen.blogspot.nl

Wagner R. Landgraf

wagner @ tmssoftware.com

Kim Madsen

www.component4developers

Peter van der Sman

[email protected]

Jeremy North

jeremy.north @ gmail.com

Benno Evers

b.evers

@everscustomtechnology.nl

Detlef Overbeek - Editor in Chief

www.blaisepascal.eu

editor @ blaisepascal.eu

Howard Page Clark

E-mail: hdpc @ talktalk.net

Rik Smit

rik @ blaisepascal.eu

Bob Swart

www.eBob42.com

Bob @ eBob42.com

Max Kleiner

www.softwareschule.ch

[email protected]

Please note: extra space characters have been deliberately added around the @ symbol in

these email addresses, which need to be removed if you use them.

editor @ blaisepascal.eu


Daniele Teti

www.danieleteti.it

[email protected]

Andrea Raimondi

John Kuiper

[email protected]

Wim Van Ingen Schenau

-Editor

wisone @ xs4all.nl

3

BLAISE PASCAL MAGAZINEis proud to announce the first edition of David Dirkses book:

COMPUTER & MATH

IN GAMES PASCAL

DAVID DIRKSEpresales at

www.blaisepascal.eu/DavidDirkse/ComputerMath_Games.html

procedure ;

var

begin

:= for to doi 1 9

begin

;end

end;

DAVID DIRKSEs

www.blaisepascal.eu/DavidDirkse/ComputerMath_Games.html

A book printed in full color, sewn back bound with a hard

cover. Quality first. A fully indexed PDF file is included.

The book contains 87 chapters, 53 projects with source

code and compiled programs (exe).

All of these projects you can download at our special

website www.blaisepascal.eu

The book is highly educational and suitable for beginners

as well as for professionals.

Play board games, solve puzzles, operate a vintage

mechanical calculator, Produce 3-dimensional computer

art, generate lists of prime numbers, explore and draw

any mathematical function.

Solve systems of equations, calculate the area of complex

polygons.

Draw lines, circles and ellipses.

Resize, rotate, compress digital images.

Design your own font, generate and reduce Truth Tables

from Boolean algebra.

And more important: understand how it all works!

For the games, winning strategies are explained.

For puzzles the search algorithm.

For all projects: the math behind is thoroughly discussed.

The Delphi 3 7 (or later) source code is available

together with full explanation. Most of the projects can

be done with FPC Lazarus as well.

Pascal is the most educative, easy to learn and only

language available for several operating systems like

Windows, Linux, Mac and Android.

It is a great programming language

COMPUTER & MATH

IN GAMES PASCAL

WATER CLOCK - CHINA - BEGINNING OF TIME (BC 4000)

Some authors claim that water clocks appeared in China

as early as 4000 BC

DELPHI

The cult of Apollo at

Delphi probably

dates back to

the 700s B . C .,

Plato in Classical Attic;

428/427 or 424/423 348/347 BC)

was a philosopher, as well as mathematician,

in Classical Greece.

Pythagoras Philosopher

Pythagoras of Samos was an

Ionian Greek philosopher, mathematician,

and founder of the religious movement

called Pythagoreanism.

Born: 571 - 495 BC,

Archimedes Mathematician

Archimedes of Syracuse was an Ancient

Greek mathematician, physicist, engineer,

inventor, and astronomer.

He is regarded as one of the leading scientists

in classical antiquity. Wikipedia

Born: 287 - 212 BC, Syracuse, Italy

Thales Philosopher

Thales of Miletus was a pre-Socratic Greek philosopher

from Miletus in Asia Minor and one of the

Seven Sages of Greece. Many, most notably Aristotle,

regard him as the first philosopher in the Greek tradition.

Born: 624 BC - 546

Euclid Mathematician

Born Mid-4th century BC - 3rd century BC

Residence Alexandria, Hellenistic Egypt

Fields Mathematics Known for

Euclidean geometry / Euclid's Elements

Euclidean algorithm

Roman Empire 700 BC

Decay of the Roman Empire 500

Building of Europe

Babbage Difference Engine No. 2

OVERVIEW OF DELPHI TO DELPHI HISTORY

Building of France 5852 BC

Building of Spain 912 and Portugal 800

Discovery of Amerca by the Vikings 990 - 1050

Discovery of America by Columbus

Columbus led his three ships - the Nina,

the Pinta and the Santa Maria -

out of the Spanish port of Palos on August 3, 1492.

Blaise Pascal (19 June 1623 19 August 1662)

was a French mathematician,

physicist, inventor, writer and Christian philosopher.

creates the first calculators

Blaise Pascal starts to gamble - result first statistics

Niklaus Wirthborn February 15, 1934 He is a Swiss

computer scientist, best known for

designing several programming languages,

including Pascal, and for pioneering several

classic topics in software engineering.

Windows

Lisa - Pascal was a Pascal implementation for the Apple Lisa workstation. It was an extension of the earlier Apple Pascal for Apple II machines,

but generated object code for 68000 processors that had to be linked

against the required libraries in the Lisa OS workshop.

Lisa Pascal laid the foundation for the development of Clascal and Mac Pascal

the first implementations of Object Pascal.

Turbo Pascal

Developer(s) Anders Hejlsberg while

working at Borland

Operating system CP/M, CP/M-86, DOS,

Windows 3.x, Macintosh

Platform 8080/Z80, 8085, x86

DELPHI XE8

DELPHI 7 released in August 2002

DELPHI released February 14, 1995

Embarcadero Technologies in 2008.

Codegear Delphi 2007.

Charles Babbage - mathematician conceived of the first programmable computer in the 1830s

Babbage never built his Difference Engine

- a mechanical calculator with thousands of parts -

because of cost overruns and political disagreements,

but the inventor passed on plans for its completion,

and in 1991, the Science Museum in London actually

built it (the printing component was finished in 2000).

As suspected, it actually works.

The roots of Turbo Pascal v1.0 started in Denmark. The first step,

in 1981, was the Blue Label Software Pascal Compiler - BLS Pascal

Compiler v1.2, copyright 1981 by Poly-Data microcenter ApS,

Strandboulvarden 63, DK 2100 Copenhagen - written by Anders

Hejlsberg for the NASCOM kit computer.

On February 8, 2006 Borland announced that it was

looking for a buyer for its IDE and database line of products,

including Delphi, to concentrate on its ALM line. On November

14, 2006 Borland transferred the development tools group to

an independent subsidiary company named CodeGear,

instead of selling it. Borland sold CodeGear to Embarcadero

Technologies in 2008. Embarcadero retained the CodeGear

division created by Borland to identify its tool and database

offerings, but identified its own database tools under the

DatabaseGear name.

Difference Engine No. 1, portion,1832

ISSUE 40

IN THIS ISSUE (43)

Issue Nr 6 2015 BLAISE PASCAL MAGAZINE 7

Euclid (300 BC), sometimes called

Euclid of Alexandria to distinguish him

from Euclid of Megara, was a Greek

mathematician, often referred to as the

"Father of Geometry". He was active in

Alexandria during the reign of Ptolemy

I (323283 BC). His Elements is one of

the most influential works in the history

of mathematics, serving as the main

textbook for teaching mathematics

(especially geometry) from the time of

Very few original references to Euclid survive, so little

is known about his life. The date, place and

circumstances of both his birth and death are unknown

and may only be estimated roughly relative to other

figures mentioned alongside him. He is rarely

mentioned by name by other Greek mathematicians

from Archimedes onward, who usually call him "the

author of Elements".The few historical references to

Euclid were written centuries after he lived, by Proclus

c. 450 AD and Pappus of Alexandria c. 320 AD

Proclus introduces Euclid only briefly in his

Commentary on the Elements. According to Proclus,

Euclid belonged to Plato's "persuasion" and brought

together the Elements, drawing on prior work by

several pupils of Plato. Proclus believes that Euclid is

not much younger than these, and that he must have

lived during the time of Ptolemy I because he was

mentioned by Archimedes (287212 BC). Although the

apparent citation of Euclid by Archimedes has been

judged to be an interpolation by later editors of his

works, it is still believed that Euclid wrote his works

before those of Archimedes.

Proclus later retells a story that, when Ptolemy I asked

if there was a shorter path to learning geometry than

Euclid's Elements, "Euclid replied there is no royal

road to geometry. In the only other key reference to

Euclid, Pappus briefly mentioned in the fourth century

that Apollonius "spent a very long time with the pupils

of Euclid at Alexandria 247222 BC.

Euclidean geometry is a mathematical system

attributed to the Alexandrian Greek mathematician

Euclid, which he described in his textbook on

geometry: the Elements. Euclid's method consists in

assuming a small set of intuitively appealing axioms,

and deducing many other propositions (theorems)

from these. Although many of Euclid's results had

been stated by earlier mathematicians, Euclid was the

first to show how these propositions could fit into a

comprehensive deductive and logical system. The

Elements begins with plane geometry, still taught in

secondary school as the first axiomatic system and the

first examples of formal proof. It goes on to the solid

geometry of three dimensions. Much of the Elements

states results of what are now called algebra and

number theory, explained in geometrical language.

For more than two thousand years, the adjective

"Euclidean" was unnecessary because no other sort of

geometry had been conceived. Euclid's axioms

seemed so intuitively obvious (with the possible

exception of the parallel postulate) that any theorem

proved from them was deemed true in an absolute,

often metaphysical, sense. Today, however, many

other self-consistent non-Euclidean geometries are

known, the first ones having been discovered in the

early 19th century. An implication of Albert Einstein's

theory of general relativity is that physical space itself

is not Euclidean, and Euclidean space is a good

approximation for it only where the gravitational field

is weak.

Euclidean geometry is an example of synthetic

geometry, in that it proceeds logically from axioms to

propositions without the use of coordinates. This is in

contrast to analytic geometry, which uses coordinates.

its publication until the late 19th or early 20th century.

In the Elements, Euclid deduced the principles of what is

now called Euclidean geometry from a small set of

axioms. Euclid also wrote works on perspective, conic

sections, spherical geometry, number theory and rigor.

Because the lack of biographical information

is unusual for the period (extensive

biographies are available for most significant

Greek mathematicians for several centuries

before and after Euclid), some researchers

have proposed that Euclid was not, in fact, a

historical character and that his works were

written by a team of mathematicians who took

the name Euclid from the historical character

Euclid of Megara.

EUCLIDES


8

AN AGE PUZZLE PAGE 1/4BY DAVID DIRKSE

On new year of the year 1997 mr. Black, a math

teacher, meets his former student White. White

remembers Blacks fascination for numbers and

greets him with: my age is equal to the sum of the

digits of my year of birth.

Black thinks for a while and then answers:

congratulations on your birthday.

QUESTIONS:

1. How Black may know that it is White's

birthday?

2. What is White's age?

SOLUTION

If Peter was born in 2000 and we live in

the year 2015 there are two possibilities:

Peter is 14 years old and his birthday

still has to come or Peter 15 years old

and his birthday is passed.

The solution may be found by checking

all possible years and calculating the

sum of the birthyear digits. Test for

digitsum = current year birth year ...(-

1 if anniversary still has to come)

This means a lot of work, so better we

write a program to do the job.

THE PROGRAM

There may be more solutions. We choose

to calculate them all and store them in a

list. Thereafter the solutions are

displayed on the screen.

For a solution we define the data type:

The solutions are displayed in a paintbox, see

picture.

We notice columns for the solution number, the

year of birth, the age and the answer of the

question anniversary passed?.

Pressing the search button starts the search

process.

Also pressing the return key after the current

year starts the search.

The text on the form is placed in Tlabel

components.

On the bottom of the form at full width there is

a statictext component for messages.

The picture below shows : 4 solutions found.

There are three reasons to make a function or a

procedure. First is the case of common code that

is needed at multiple places of the main program.

This reduces the total code.

The second reason is clearity. We place specific

code apart for readability.

For this reason we made a function to sum the

birthyear digits.

And these global variables:

Thisyear comes from a TEdit component

(this year) where the current year was typed.

Solutions is an array[1..maxsolution] of type

TSolution.

SolutionNr is the number of stored solutions.

This value is 0 if no solution exists.

Maxsolution is a constant set at 20.

Tsolution = record

: ;birthyear word

: ;age byte

: ;birthdaypassed boolean

;end

var : ;thisyear word

: [ ] ;solutions maxsolution Tsolutionarray of1..

: ; solutionNr byte

9MaxSolution is a constant set to 20, the space in

the solutions array.

Which saves typing. Otherwise we had to add

solutions[solutionNr]. before

birthyear, age. Now this is done by Delphi.

The procedure ShowSolutions:

function ( : ) : ;sumdigits year word byte

var string : ;s

: ;i byte

begin

:= ( );s inttostr year

:= ;result 0

:= ( ) := + ( [ ])- ( );for to doi length s result result ord s i ord1 '0'

end;

Ord('0') is 48, the ASCII code of digit 0.

So ord('8') ord('0') = 8.

Without -ord('0') we would get 56.

THE SEARCH PROCESS

In procedure searchBrtClick, called by the search

button, we notice following local variables:

var : ;digitsum byte

: ;birthyear word

variabele birthyear runs from

startyear...thisyear.

startyear is a constant set to 1900.

Digitsum is the sum of the digits of

birthyear.

Before the search starts two checks are made:

1. the current year must be defined

2. the current year must be equal or bigger than

startyear.

Then solutionNr is set to zero.

All values of birthyear are checked.

If the test yields true, age (=digitsum) and

anniversary passed is added to the solutions

array. This program loop looks like:

for to do := birthyear startyear thisyear

begin

:= ( );digitsum sumdigits birthyear

= - if thendigitsum thisyear birthyear

( , , );savesolution digitsum birthyear true

= - - if thendigitsum thisyear birthyear 1

( , , );savesolution digitsum birthyear false

; end //for

The procedure SaveSolution:

procedure ( , : ; : );SaveSolution ag by word bp boolean

//ag:age; by:birthyear; bp:birthday passed

begin

( );inc solutionNr

Issue Nr 6 2015 BLAISE PASCAL MAGAZINE10

Why all this preset constants? Why not using the

values directly? Sure we could. But using

constants is smarter because they make future

changes much more easy. Say we want to change

the startyear. Without the constant definition we

would have to change all places where the

constant (1900) was used. But defined as a

constant we only need to change the constant

value once to be effective in all places in the

program.

PROCEDURE BEFORE CALLED

CLEARPAINTBOX.

This procedure erases the paintbox by painting

the canvas white. (for this color we could have used a

constant as well)

procedure ;clearpaintbox

begin

. with doform1 PaintBox1

with docanvas

begin

. := ;brush Style bsSolid

. := ;brush color $ffffff

( ( , , , ));fillrect rect width height0 0

;end

end;

The paintbox has properties width, height and

canvas. By not using with

form1.paintbox1.canvas but the separation

with form1.paintbox1 do with canvas do, we may

use both the canvas and the paintbox properties

without the name prefixes. Canvas property

brush takes care of background coloring.

Style is a property of the brush. If we write

brush.style := bsClear the fillrect does nothing,

the brush is turned off. Fillrect( r ) paints

rectangle r of the paintbox in the color

brush.color. r is of type Trect. Function rect

(defined inside Delphi) makes a Trect from

coordinates.

See picture:

We notice that the width occupies left ...

right-1 and the height is from top to

bottom-1. Width = right left.

Height = bottom top.

SETTINGS

In the TEdit component YearEdit we enter

the current year. Property maxlength of

YearEdit is set to 4. This prevents entering

more than four digits.

The event YearEditKeyPress points to a

procedure of this name. Reason is to allow only

the digits 0..9 together with backspace for

procedure . ( : ; TForm1 yeareditKeyPress Sender TObject

var : );Key Char

begin

( [ .. , ]) := ;if not in thenkey key'0' '9' #08 #0

end;

8 is the ASCII code of the backspace key.

Prefix # denotes a following character code.

The data between [] is of the SET datatype.

We want to start the search as well by pressing

the return key after the current year.

The return code is catched before it reaches the

Tedit. This is done by setting the Tform

property KeyPreview to true and defining the

event FormKeyDown.

procedure . ( : ; TForm1 FormKeyDown Sender TObject

var : ; : );Key Word Shift TShiftState

// starts search

begin

= if thenkey VK_RETURN

begin

:= ;key 0

( );searchBtnClick self

;end

end;

VK_RETURN is a predifined Delphi constant

representing the return key.

NOTE: key now is of the type word, not char.

Procedure searchBtnClick is called

with Form1 (self) as sender.

For more details please refer to the source code.

FINALLY

With the help of this program the reader may

solve the puzzle without doing the arithemetic

himself. Now it becomes clear why Black is sure

today is White's birthday.

Other, more difficult, questions would be birth

years that have zero or multiple solutions.

This would require some extensions of the

program.

Have fun!

AN AGE PUZZLE PAGE 3/4

50 hands-on recipes to master the power of Delphi for

cross-platform and mobile development on Windows,

Mac OS X, Android, and iOS

Daniele Teti

Quick answers to common problems

Delphi Cookbook

See our special offer:

if you take out a subscription

for two years the book

will cost you only 10,00

http://www.blaisepascal.eu/daniele_teti_book/DanieleTeti.html

OICH EC SR

OT

ID

E

A

MAZIN

G

BLAISE PASCAL

MAGAZINE

30,00

including VAT

39

including

the

printed

book,

ebook

and shipping


BY BOIAN MITOV

In the previous articles, you learned how to

program Arduino using Visuino, and how to

communicate with it using USB simulated serial

port from your Delphi code. This opens a lot of

interesting possibilities for collecting and

processing live data, but the direct USB

connection imposes some limitations. What if you

want to collect data from many sensors spread

over large area? Or what if you want to

communicate with remote sensors over Internet?

The basic Arduino UNO does not have built in

network adapter, but there is Ethernet shield

available for it. In addition many of the more

advanced Arduino boards and their clones come

with WiFi or wired Ethernet built in.

There are also cheap and simple ESP8266 WiFi

modules that can be connected to the Arduino, so

networking Arduinos is routinely done.

In this article you will learn how to setup Arduino

to use Ethernet Shield, how to program it with

Visuino, and how to connect to it from a Delphi

application over the local network or Internet.

Before you start, you will need to install Ethernet

Shield on the Arduino. This is fairly easy. Just

snap it on top of the board as shown in the

picture.

Add Ethernet Shield:

Next you need to specify the MAC address for

the shield. You can use a MAC address generator,

or one of the MAC addresses from the Arduino

tutorials. Here I use DE-AD-BE-EF-FE-ED:

You also will need to install CommunicationLab,

PlotLab and InstrumentLab from Mitov

Software. CommunicationLab is not officially

released yet, but prerelease builds are available

on request. You can also easily modify the

examples in this article not to use PlotLab or

InstrumentLab.

First you will create a simple Arduino server.

Start Visuino. Click on the Down arrow button in

the top right corner of the Arduino component,

and from the menu select Add Shields... :

ARDUINO: THE VISUINO PROJECT - PART 4 PAGE - 1/15

INTERNET OF THINGS WITH ARDUINO AND DELPHI

13

Set the IP address for the Arduino

as example 192.168.0.55:

And set the Enabled property of the IPAddress

to True so the IP address will be used when

Arduino starts:

Once the Ethernet Shield is configured, you can

add one or more TCP/IP Client, TCP/IP Server,

or UDP sockets to it. Click on the ... button

after the Sockets to add a socket:


Add TCP/IP Server socket:

Set the UseDHCP to false, so Arduino will

work with a fixed IP address:



6 Issue Nr 6 2015 BLAISE PASCAL MAGAZINE14

Set the Socket Port property to 8080:

To generate some test data from Arduino, you

can use a Sine Generator as shown here, or

you can use any other source of Analog data,

or one of the Analog channels:




Connect the data source to the Input Pin of the

Server Socket:

Press F9 to generate the Arduino

Sketch and open the Arduino IDE:




Click on the Upload button to compile and

upload the sketch:

The simplest way to see if the Arduino project is

working is to open a web browser and enter the

Arduino IP address, and socket number

192.168.0.55:8080 . You should see the data

appearing in the browser, in this case Chrome:



17Issue Nr 6 2015 BLAISE PASCAL MAGAZINE

Next its time to receive the data in Delphi.

Start RAD Studio, create a VCL Form project,

and drop a TCLClientSocket from

CommunicationLab on the form:

Set the IP Address to the same used in the

Visuino project 192.168.0.55: Set the Port to 8080:



18 Issue Nr 6 2015 BLAISE PASCAL MAGAZINE

Switch to the OpenWire tab, and connect the

Output Pin of the Socket to the Input Pin of the

Terminal:

Compile and run the application. You will see

the data arriving in the terminal:

Drop TCLTerminal on the form:




If you need to access the data in your code,

the Socket component has OnReceive event:

Now you can receive data from Arduino over the

network, however the data arrives from a single

sensor, and in text form. This makes it difficult

to work with, and limits the data channels that

we can get.

As shown in the previous article, Visuino and

CommunicationLab have support for packet

data. You can use it with sockets the same way

you did with the serial port.

Start a new Visuino project, add and configure

the shield and the socket as you did in the

previous project.




Connect the Output Pin of the Packet component

to the Input Pin of the Socket:

Double click on the packet component to open

the elements editor. In the editor add 2 Analog

and 2 Digital elements:

Next, add a Packet component:




Connect the Input Pins of the Analog Channels

to the Output Pins of Analog Input Channel[ 0

] and Analog Input Channel[ 1 ], and the

Input Pins of the Digital Channels to the Output

Pins of Digital Channel 0 and 1 of the Arduino

Board component:

Expand the HeadMarker and for the Bytes click

on the ... button:




In the Bytes editor enter 55 55, then click OK:

This will enure that the packet has unique

header and its starting point can be identified in

the data stream. Visuino uses special algorithm

to ensure that the header marker will not appear

in the packet itself.

Press F9 to generate, then compile and upload

the sketch in the Arduino IDE as you did in the

previous project.

Now that the Arduino is ready, lets switch to

Delphi.

Start a new VCL Form project, add the

TCLClientSocket, and set the IP Address and Port

as in the previous project:




Next add a TCLUnpacket, TSLScope,

TILAngularGauge and 2 TILLed components:

Switch to the OpenWire tab, and double click on the

CLUnpacket1:



24

Add 2 Float and 2 Boolean channels:

Expand the HeaderMarker and for the bytes click

on the ... (called ellipsis)





In the Bytes editor enter 55 55, then click OK:

Connect the components as shown in the picture:




Compile and run the application. You will see the

data arriving from Arduino over the 4 channels:

When you need to access the data from code,

you can use the TSLGenericRealValue as

example to receive the Floating point data and

process it in the OnProcessData event, as

shown in one of the previous articles. There are

also similar components for processing the

Boolean data included in LogicLab.

The communication to Arduino is equally easy.

In order to send the data from Delphi, use

TCLPacket component and in Visuino use

Unpacket component.

BLAISE PASCAL MAGAZINE subscribers that

visit our PASCON - Event will receive a DVD with

lots ofprograms, information and as a

VERY SPECIAL INCENTIVE you will get an

ARDUINO-BOARD FOR FREE INCLUDING

THE VISUINO SOFTWARE from Boian Mitov to

be able to compose and create your own

software for the board

CONCLUSION

This article has given you enough information to

start communicating with one or more Arduino

devices over wired network, or Internet from your

Delphi code. This is your first introduction to the

exciting world of Internet of Things. In the

following articles you will learn how to

communicate with Arduino over WiFi, and how to

make multiple Arduino boards to talk to each

other.



The components found in the Visuino software represent their hardware components and you

will easily be able to create and design your programs using drag and drop. No equipment or

hardware is needed to run the software in design mode. Once you have completed the design,

you can connect Arduino board upload and run it.

For those people who are not strong on writing code then designing, compiling and creating

Arduino programs has never been easier! Why waste time on creating code when we have done

all the hard work for you already? You have your Arduino board, and great hardware design,

see it running in minutes, not hours!

Currently we are running a Beta program which you can be part of by joining our Google group.

Join the group now to download and test the software or send an email to [email protected].

AVAILABLE ON THE NEXT

PASCON 15 SEPTEMBER 2015

What is Visuino?Visuino is the latest innovative software from Mitov Software. A visual programming

environment allowing you to program your Arduino boards. Although it currently

supports the official Arduino boards, it is not restricted to their support alone and

requests to support new hardware are welcome.

INTRODUCTION What is Arduino?

VISUINO IS THE LATEST INNOVATIVE PRODUCT

FROM MITOV SOFTWARE.

www.visuino.com


DATABASE WORKBENCH 5 PAGE 1/8 THE SWISS ARMY KNIFE FOR DATABASES BY PETER VAN DER SMAN

When you start working with databases sooner or

later you'll come at a point where you start

searching for a tool to maintain your database.

After a bit of searching you'll end up using

FlameRobin when your database is a Firebird

database, using other databases might result in

other programs. This is cumbersome at the

moment you can't avoid using different databases.

For instance when you have two customers each

using their own database. Database Workbench is

database tool working independent from

databases, that is to say it supports a lot of

different database managers. Time to take a closer

look at it.

EXAMINING A EXISTING DATABASE

As stated before, Database Workbench does

support different databases, but is should be said

the databases you can use is subject to the licence

you have. Having said that we start using

Firebird as a starting point as this seems to be the

database propagated for using with the current

versions of Delphi. So we start our freshly

installed program and firstly we have to select

our previously installed Firebird as Database

server to use. Then we can start opening a

existing database. As an example we choose the

Employees.fdb which comes as a demo with

your Firebird installation. It gives you directly a

nice overview ( right column )Figure 1. The content for Employees.fdb

in the Navigator

Of course we can use the Navigator as above to

study the content of all the tables in the database.

But Database Workbench offers a much more

handy option for this. After opening a Diagram

Editor we can use the option Reverse Engineer

Database to build a schematic overview for the

database. We can even select to show only a part

of the Database but, greedy as we are, we of

course select for the whole database.

Then we can give the things we like a closer view.

In the bottom part we find all kinds of

information about the database itself. A nice one

are the Connections under Activity. Opening

this gives a screen showing which programs are

currently connected to the database. Of course it

will allways show at least one connection as the

program itself is connected to it. After making

another connection, for example in the Delphi

Data explorer, a second entry will show up (in

this case bds.exe). Always good to know in case

you're wondering who is blocking the database.


Figure 2. Employees.fdb in the Diagram Editor (part)

Always difficult to get a nicely looking schema

for all this linked tables. You should consider it

as a starting point to make a schema as nice as

you want to have it. You can replace all tables,

smarten the links, whatever you like. The

Navigator is a nice tool to show where you are

in the total schema. Or to navigate quickly to

another part. After some manual adjustments

our schema could be like figure 3.

DATABASE WORKBENCH 5THE SWISS ARMY KNIFE FOR DATABASES PAGE 2/8


Figure 3. Employees.fdb in the Diagram Editor

(after manual adjustment)

Who would have thought this is all needed for a

sample employee file! This schema is just of a

size to overview it at a glance. In the real world

your schema will easily by a lot larger. Luckily

enough we have the possibility to make a Sub-

diagram so you can divide your schema in

relevant parts.

In the schema we directly see which fields are

part of the primary key (having a golden key)

and which play a role in a reference (having a

grey key). More information about the table we

get with a double click on it. For example:

Figure 4. Columns for table EMPLOYEE


31

More than enough information. You might

notice the Column Type for EMPNO for the

column EMP_NO. Looks a bit odd, this has to

do with defining your own types in Firebird.

In the schema (figure 1) you can find these

definitions under Domains. For our database

has no less than 15 types defined.

Our EMPNO' turnes out to be of type

SmallInt. The other tabs in the Table

Properties form show more information about

the table. The tab DLL gives you all the code

needed to create the table.

This design is used consequently. Double

clicking in the Navigator (figure 1) on

different items will result in similar screens to

popup. No matter if it is about a Domain, a

View, an Index, a Constraint, a Trigger,

an Exception or any other item, you'll get Tab

with relevant detail, and a tab DDL giving

you the code to implement in the database. In

fact, you don't really need this code, as we'll see

later on, but it can be a handy way to check up

what you have done with an example

implementation. Unless you're the type

inventing all by yourself and never looking at

example code?

The temptation is strong to directly start making

all kind of changes. The schema (figure 3) shows

us an emp_NO, proj_ID en job_CODE. Of

course there can be puristical reasons the make

the distinction between NO, ID and

CODE, but perhaps you would prefer to

change this all to ID. The good news is we

actually can easily make these changes in the

screen as shown in figure 4. Just change

EMP_NO to EMP_ID. The bad news is it

won't be of any use.

The diagram in the diagram editor is not

connected to a physical database.

The consequence is we cannot pass our changes

made to the database. In fact this is a good thing.

In a designing phase it could be a nice idea to

make the proposed changes, making them in an

existing database can be quite disastrous.

There's every change that these fields play a part

in a trigger, a view or a procedure. In fact, we

can change fieldnames using Database

Workbench: just open a table from the navigator

and change the name of the field. In this screen

there is this handy button Create/Alter Table.

Nine times out of 10 this will result in a message,

from Firebird, telling you the update was

unsuccessful due to data integrity.

But changing the AGED field from the table

SALES will work fine using this option.

The other way around will work fine. The option

Update From Database can be used to match

your diagram with an existing database. The

lay-out we made with so much effort (figure 3)

will be maintained.

DESIGNING A NEW DATABASE

The previous part showed Database Workbench

5 can be used to analyse the structure of a

database we received from a third party. But is it

of any good designing a new database? You

understand this question is rhetorical. Of course

we can. Let's just do it!

CONCEPTUAL AND PHYSICAL DIAGRAMS

The best way to start a new design is designing

it using the Diagram Editor. Starting with an

empty Diagram we'll find under the New

button two options: New Conceptual

Diagram and New Physical Diagram. It is

just like is sounds: you can start making a more

sketchy diagram or directly start with a design

tailored to the possibilities your database

manager of choice provides. If this would not be

clear at first sight: after selecting a New

Physical Diagram you must point out for which

database you're going to design.

Figure 5. The first conceptual

diagram

So we choose for a

Conceptual

Diagram and get an

empty working area

where we can drop

new tables. Of

course I could have

chosen to redesign

this inevitable

databases like

Employees or

FishFacts, but let's

design a brandnew

database we can use

to browse all articles

ever published in

Blaise. Our first go

would be to have

three tables: blaise

edition, writer

and article. So we

drop three tables (in

this phase named

Entities) on our working area, double click on

them to give them meaningful names and

adding some fields (in this phase named

Attributes). After some rearrangement it could

look like this




Note I did'nt mind about the type definition

for my fields. No reason to bother about it

right now, we can just focus in getting the

general structure right.

Looking to our design we note the field

Language appearing in two tables. So

most likely we should consider adding a

new table/entity Languages. And of

course we must find a way to bind the

Articles to the proper Blaise edition.

The most flexible way to do it is using a new

table making the connecting between the

two tables. It gives the editor a way to

republish an article . Speaking about this:

most articles in effect will be republished in

different editions, that is to say, in

translation. So apart from the writer of the

original version there can be a translator

too. And we should be able to refer to the

original article. So we can add a table

Translators, but the easy way is to use the

table Writers for this purpose as well. This

directly solves the problem of persons who

both write articles and translate articles as

well. And besides, writing an article or

translating, both are a kind of writing. And,

in case you had noticed, we now have a

reason to use author in the Articles table

Figure 6. The half worked out conceptual diagram

In the real world this would be the moment to

put it aside and to look at it again at a later point

of time and discussing it with a colleague.

However, in the world of Blaise the next edition

is coming soon, so we continue with our schema

right away. We do a bluff and ask Database

Workbench to convert this schema to a

Physical Diagram for Firebird. The result is a

pile of error messages: this is the time we should

point out the datatypes we are going to use.

Further on we did make connections between

tables, but I didn't really bother working them

out. The good news is we don't have to search

ourselves for missing data, we just get a list of

items to resolve. We can't blame the program it

initially is a long list. The following schema

shows the more adapted conceptual schema

(figure 7).

referring to a table writers. And a fieldname

date is a bit ambiguous, so let's change it to

DatePublished right away.

Let's make some connections as well. The icons

as shown in figure 5 point out we can use three

kinds of links: Identifying Relationship,

Non-Identifying Relationship and Documentation

link. The last one is meant to make connections

just for clarifications in the diagram, not for

implementing in the database. After some not

too hard labour it could look like this.

Figure 7. The fully adapted conceptual diagram



Where needed I provided for a unique key field

in all tables. All fields are given a type. Just for

showing the differences I used some different

types. And the connections are filled in with

relevant information. To get a nice picture I

removed the connection between

OriginalVersion and ArticleID from the

Diagram. But it still exists. To show it a Sub-

diagram can be used:

Figure 7b. The Articles sub-diagram

FROM CONCEPTUAL TO PHYSICAL DIAGRAM

So now we are ready to convert to a physical

schema. This is the point to decide on which

database manager our database must be build. The

next thing to select the option Generate Physical

Diagram and select for Firebird resulting to the

following screen.

Figure 8. Conversion to Firebird

We just take the standard options, including

Firebird 3.0, and then click OK. That's all to get a

diagram adjusted for the desired database. All

types are translated to types provided by our

database. To illustrate this I made two

conversions, one to Firebird and one to MySQL.

The differences for the table Articles are

shown in figure 9. In fact you'll see just one

difference, the conversion of the type Unicode

text in the Abstract field . For Firebird it is

converted to BLOB(text), for MySQL the

LongText is used.

Figure 9a. Table Articles for Firebird.

Figure 9b. Table Articles for MySQL

After looking closely to the final schema

again, it will probably result in some

desired changes. We don't have to go back

to the conceptual diagram, we just can

make the changes in our conceptual

diagram. The interface is nearly the same.

So - if this is the moment - we realise the

Dutch Edition 126 is a translated version

of the English 44 edition and we should

provide a way bring this into the database

we can adjust the table Blaise (in the

same way we did for Articles):

Figure 10. Adapted Table Blaise


34

The reference from OriginalVersion back to the

table Blaise is called a Constraint. It is build

using the following screen.

Figure 11a. Definition of a Constraint

And, as promised before, if we really would be

interested we could find the code needed to add

this Constraint to the database in the DDL tab:




Of course now we have to doubel check the total

schema again and adapt it where necessary.

We could add a Check Constraint, an Index or a

Trigger where desired.

After doing this we can use the schema to build a

real database. We have two options available.

The hard way is to use the option Extract DDL For

ALL Objects. It provides us a SQL script to be

used for generating the database. The advantage

might be you can recheck the script before

running it and/or maybe add some very special

code you couldn't add using the program.

And the script can be useful for documentation

purposes. But we also can use the easy way.

Just take the option Create Database from

Diagram and follow the instructions in the

Wizard. It will ask you the location of the

database, username and password and so on.

When we're done the database will be created

and the script will run in order to create your

structured tables. In anything does go wrong

with the script it will pop-up in the script editor

so you can make your adjustments. .


EPILOGUE

Database Workbench is a handy tool for

working with databases. Either in case you want

to examine an unknown database delivered to

you or in case you want to design a brand new

database, in both cases the program offers you a

lot of support to do so. Of course you can find

loads of other programs doing this. The nice

thing about Database Workbench is it supports

multiple databases. So in case you're working

with different databases you only have to know

about one single tool instead of learning about a

tool for each different database

36 Issue Nr 6 2015 BLAISE PASCAL MAGAZINE

better office benelux | asterlaan 6 5582EH waalre | 040 222 26 43 [email protected]

Benno Evers is our specialist for questions about

kbmMW.

He can help you with basic questions regarding kbmMW

as well as with turnkey Development and Consultancy.

Hes a specialist in netwoks, internet and hardware.

Specialist help and consultancy

for kbmMW

[email protected]

CUSTOM TECHNOLOGY

COMPONENTS

DEVELOPERS4

http://www.blaisepascal.eu/DucthPascon/Pascon2015UK.html

1 2 34

5126

7 8

1314 15

16 17

25 262720

2128 22 23

24

3334 35

36

37 38

39

40 41

42

3229 30 31

910

1118 19

WINDOWS 10

L I B R A R Y 2 0 1 5

B L A I S E P A S C A L M A G A Z I N E

A L L I S S U E S I N O N E F I L EA U T H O R S A L FA B E T I C A L

CLICK TO OPEN

ITEM (NUMBER)

CLICK TO OPEN

ITEM (NUMBER)

CLICK TO OPEN

DELPHI URL

CLICK TO OPEN

A BROWSER WITH

URL

CLICK TO OPEN

THE TOTAL

OVERVIEW OF ALL

ITEMS IN ONE

CLICK TO OPEN

THE TOTAL

OVERVIEW OF ALL

ITEMS IN ONE

www.blaisepascal.eu/subscribers/UK/UK_CD_DVD_USB_Department.html


Delphi

expertstarter

Issue Nr 6 2015 BLAISE PASCAL MAGAZINECOMPONENTS

DEVELOPERS439

TIPS AND TRICKS WITH KBMMEMTABLE PAGE 1/9

kbmMemTable is an in memory row/column

oriented database which have a vast

number of features and which is very fast in

both storing data and locating stored data.

I will in this article focus on some basic

things and on some of the less known,

but very useful, features of kbmMemTable.

var

: ;mt TkbmMemTable

begin

// Define fields (you may already

// have them defined at designtime).

. . ( , );mt FieldDefs Add ftInteger'ID'

. . ( , , );mt FieldDefs Add ftString'Name' 80

. . ( , , );mt FieldDefs Add ftString'Address' 80

// Define indexes (if you have a large amount of records

in your mt

// indexes will make searches faster, but inserts/edits slower).

. . ( , ,[ ]);mt IndexDefs Add ixPrimary'iID' 'ID'

. . ( ,mt IndexDefs Add 'iName'

Name',[ ]);ixCaseInsensitive

. ;mt CreateTable

. ;mt Open

end;

When searching on for records using the Name

field as search criteria, we have defined that it

should search case insensitively, so for example

uppercase A is the same as lowercase a.

FAST INSERTION INTO THE DATABASE

kbmMemTable exists in two primary versions,

Standard Edition and Professional Edition.

Standard Edition can be purchased separately,

while Professional Edition only is available as a

bundle with kbmMW Professional or kbmMW

Enterprise Edition.

While kbmMemTable Professional is the

absolutely fastest memory table in the world,

kbmMemTable Standard Edition can almost reach

its performance when used correctly.

If you are to insert a large number of records in

kbmMemTable Standard Edition, then you will

want to do like this:

var

,fldID

,fldName

: ;fldAddress TField

begin

// First get quick access to the fields.

// This prevents having to search for a field each

// and every time a record is to be inserted.

:= . (' ');fldID mt FieldByName ID

:= . (' ');fldName mt FieldByName Name

:= . (' ');fldAddress mt FieldByName Address

// Prevent update of attached controls on each insert.

// This is a big performance factor when inserting huge

// amounts of records.

. ;mt DisableControls

// Prevent update of indexes on each insert. Professional

// Edition is much faster in on the fly updates of indexes,

// so it will perform extremely fast even without disabling

// indexes while inserting. But the highest performance

// is obtained by disabling them before the batch insert.

. := ;mt EnableIndexes false

// Insert a lot of records.

:= for to doi 1 1000000

begin

. ;mt Append

. := ;fldID AsInteger i

. :=' '+ ( );fldName AsString Name inttostr i

. :=' '+ ( );fldAddress AsString Address inttostr i

. ;mt Post

;end

// Enable and update indexes.

. := ;mt EnableIndexes true

. ;mt UpdateIndexes

// Enable updating attached controls (grids etc).

. ;Mt EnableControls

HOW TO SEARCH?

You either choose to switch to the index

representing the column(s) you want to search

on, or you simply just search and let

kbmMemTable find a relevant index (if any) to

use for the search.

If you want the absolutely fastest result, then you

should switch to the index first, to avoid that

kbmMemTable have to make an additional

search on your current index to sync with the

found place in the search index.

However the second index sync search is

generally very fast.

You can use Locate to search for a complete name

for example. If it finds a record with the given

name, that record will be the current record, and

you can access other fields in it immediately.

if then . ( , ,[]) mt Locate 'Name' 'Jens Hansen'

(MessageDlg 'Found record.

Address=' 'Address'+ . ( ). );mt FieldByName AsString

CREATE A MEMORY TABLE WITH

INDEXES IN CODE

The Codegear edition is for free and can do it all

But it only supports specific versions... eg only

new xe8 has supprt for latest kbmmemtable.

Xe7 supports older kbmmt and edition xe2 even

older etc.

BY KIM MADSEN

PAGE - 2/4

Issue Nr 6 2015 BLAISE PASCAL MAGAZINE COMPONENTS

DEVELOPERS440

You can use Lookup to search for a record

containing the given value, and return the

contents of any field as result. It will not move

the current record.

v mt Lookup:= . ( , , );'Name' 'Jens Hansen' 'Address'

( ) if not thenVarIsNull v

( + );MessageDlg v'Found record.Address='

You can also search for partial values

if . ( , ,[ ]) mt Locate loPartialKey'Name' 'Jens'

then MessageDlg mt FieldByName AsString( + . ( ). );'Found record. Address=' 'Address'

How to search on multiple fields?

Let kMemTable automatically create indexes for you

to improve search performance

This may however result in creation of many indexes

over time if you often search on new columns and

having many indexes to maintain when

inserting/editing/deleting records will always give a

performance penalty, so use with care.

HOW TO SORT?

Adding an index will always result in the data

being sorted according to an index. So you can

add an index and switch to it.

Or even easier:

HOW TO ONLY SHOW CERTAIN RECORDS?

For this purpose, you can use a range or a filter.

A range can limit the records displayed in for

example a grid, to show only records within a

range (eg. 10 Greater than

< Less than

>= Greather than or equal

Issue Nr 6 2015 BLAISE PASCAL MAGAZINECOMPONENTS

DEVELOPERS441

This creates a new index, ordered (and quickly

searchable) by ID, and filtered by the given

filter string. You can filter on any field or

calculation of fields. It doesn't have to be fields

that are in the key fields list.

Every time you add or alters records, the index

will be updated and records may this way

disappear or appear in the filtered index

depending on the values of the record.

When you switch to a filtered index, and thus

only show the records participating in that

index, scrolling through or searching isjust as

fast as if you had no filter defined. Inserts and

updates have a small performance penalty, so

if you have many records to insert, use the

batch insert method outlined earlier in this

article.

STANDARD INDEXES?

With standard indexes, I refer to indexes

which are created/defined by the memtable

during its course of operation.

When a memory table is opened, there always

exist one single index, the row order index. It's

the index that is updated to ensure that order

of insertion of records can be established.

It's also responsible for holding a reference

to all records. The row order index is selected

by setting the IndexName to an empty string.

Internally you will find the roworder index to

be named '__MT__ROWORDER_'. A string constant

exists for that name: kbmRowOrderIndex.

The row order index can't be deleted.

Further a number of other indexes may be

defined at various occations.

Functions:

UPPER( expression )

Convert string expression to uppercase.

LOWER( expression )

Convert string expression to lowercase.

SUBSTRING( expression , startpos , count )

Extract a substring from the expression string

starting at startpos and with max length of count

chars. The count parameter can be omitted, in

which case is means the rest of the string.

TRIM( expression ) Trim string expression for all

leading and trailing spaces.

TRIMLEFT( expression ) Trim string expression

for all leading spaces.

TRIMRIGHT( expression ) Trim string expression

for all trailing spaces.

GETDATE Returns a floating point date/time

value for current time.

YEAR( expression ) Return the year of a floating

point date/time value.

MONTH( expression ) Return the month of the year

(1..12) of a floating point date/time value.

DAY( expression ) Return the day of the month

(1..28/29/30/31) of a floating point date/time

value.

HOUR( expression ) Return the hour (0..23) of a

floating point date/time value.

MINUTE( expression ) Return the minute (0..59) of

a floating point date/time value.

SECOND( expression ) Return the second (0..59) of

a floating point date/time value.

DATE( datestring , formatstring )

Convert the datestring to a floating point

date/time value according to the

SysUtils.FormatDateTime format string .

DATE( expression ) Return the date part of a


TIME( timestring , formatstring )

Convert the timestring to a floating point

date/time value according to the

SysUtils.FormatDateTimeformat string .

TIME( expression ) Return the time part of a


Filtering however is done on each record every

time that record is accessed, thru scrolling or

counting number of records etc. So it's fairly slow

on large datasets to use a filter.

A better way, is to create a filtered index. That is a

real index which only contains references to records

that live up to the filter expression.

mt AddFilteredIndex. ( , ,[], ,[]);'iFilteredIndex' 'ID' '(ID>=10) AND (ID


DEVELOPERS442

kbmMemTable supports whats called

transaction management via the methods

StartTransaction, Commit and Rollback.

Further it also supports an Undo function on

the record level. More about that in a moment.

To use transaction management, one also need

to understand versioning.

kbmMemTable have two properties controlling

versioning: EnableVersioning and

VersioningMode. EnableVersioning is a

Boolean, while VersioningMode can have

the value of mtvm1SinceCheckPoint or

mtvmAllSinceCheckPoint.

Versioning allows kbmMemTable to store

multiple versions of the same record.

Whenever a record is inserted into the

memory table, one version of that record

exists, the one that was inserted.

Default EnableVersioning is false and thus

the records you see in the memory table are

the ones you have.

No other versions are created.

If EnableVersioning is set to true, then

kbmMemTable will start to make multiple

versions of a record, depending on the setting

of VersioningMode.

If VersioningMode is set to

mtvm1SinceCheckPoint, then at most, an

original version is kept, and the altered one.

This is cool, if you need to update an external

storage with the changes made in the memory

table.

Then the original version identifies what you

have to search for, and the new version

identifies what to change to.

A deltahandler is what is used for handling

those scenarios, but now we will focus on

transactions. So in this case, you can only

undo back to the original version, and not

intermediate changes.

INDEX INHERITANCE

If you want to temporarily sort on a different

field than the ones participating on the current

selected index (it could for example be a filtered

index), you simply do SortOn, as described

earlier on, while having the filtered index

selected.

This will create a new sorted index, based on

the filtered index, so only records that are

acceptable in the base index, are candidate for

inclusion in the new sort index.

TRANSACTIONS AND VERSIONING

If you make modifications to a number of

records in a table, it would, in some situations,

be nice to be able to roll back those changes in

one go. For example if you have an application

where the user can make modifications to a

number of entries (inventory for example), and at

a later stage decide to save those changes or

cancel them. Cancelling the changes would

require a roll back, while saving means that the

user commits to his changes.

'__MT__DETAIL_' (kbmDetailIndex).

This index is created (and automatically recreated)

when the master record changes in a master/detail

relationship. It ensures that browsing through

detail records is very fast, and quick to search on.

'__MT__DEFSORT_' (kbmDefSortIndex).

This index is created when you are using the Sort

or SortOn methods. It is only recreated when you

issue another Sort/SortOn method call.

'__MT__DEFAULT_' (kbmDefaultIndex).

This index is automatically created when you set

IndexFieldNames to a set of fields for which no

index already exist.

'__MT__RANGE_' (kbmRangeIndex).

This index is automatically created when you

apply a range filter.

'__MT__AUTO_' (kbmAutoIndex). This is

actually not only one index, but potentially many

indexes, which are automatically created if you

have the property AutoAddIndexes set to true,

and you make a search locate/lookup)

on fields for which there are no existing indexes

available. The actual indexname will be

'__MT__AUTO_' + the semicolon separated field

names. Be aware that this index is not

automatically destroyed, so if you search on many

weird combinations of fields, you may end up

with many indexes, and may need to delete them

yourself if needed.


If you compile and upload the Arduino code as described in the

previous articles, and then connect to it with Visuino, you will see the

sine wave deformed by the quadratic function plotted in the scope:


THE DELTAHANDLER

A deltahandler is a class that can scan through

a memory table, and determine what has

happened with each record, and allow the

developer to do something depending on what

happened.

Was it inserted, deleted, modified or was

nothing changed with it?

It is essentially a great tool for making changes

in your dataset resolve back to from where the

data originated from (typically a SQL database or

files in a file system etc).

This is a simple deltahandler:

Example:

In this example a record is inserted, a transaction is

started, then the record is edited and then deleted.

With VersioningMode set to

mtvm1SinceCheckPoint, then you can only

choose to roll back to where the record was

inserted. The version containing the edited record is

lost the moment the record is deleted. Only one

version in addition to the original record is kept.

To allow for a multi level undo, then

VersioningMode should be set to

mtvmAllSinceCheckPoint. Then you can do

this:

You see you can nest transactions this way.

If you want to make a change, that is protected by a

StartTransaction, stick, then you call

mt.Commit. After the commit, the change can only

be rolled back if there are multiple levels of nested

transactions as in this example:

You can check the state of a record by checking the

UpdateStatus property. It returns the current

records state. Was it inserted (usInserted),

deleted (usDeleted), modified (usModified)

and unchanged (usUnmodified). To actually see

usDeleted marked records, it require adding and

switching to a special index that allows for showing

deleted records check the overloaded version of

AddIndex which accepts providing a set of

UpdateStatus flags, for which records to include

in the index.

When you insert/append records, those will be

marked with usInserted as their

UpdateStatus propery. That also happens when

you load the records from another dataset, because

technically they have been inserted into the

memory table. You can however ask kbmMemTable

to consider these newly loaded records, as being

the original ones and thus that they should have

the UpdateStatus of usUnmodified.

The latest version of any record is now considered

the original version and marked as

usUnmodified, and all intermediate versions

have been removed. That also means that if records

were deleted, they are now permanently deleted

and can't be recovered. To undo last change of the

current record, you can use:

Remember that the number of undo's you can

make on the current record, depends on the

setting of EnableVersioning and

VersioningMode.

// An example on how to create a deltahandler.

TMyDeltaHandler TkbmCustomDeltaHandler = ( )class

protected

( : ; procedure varInsertRecord Retry boolean

var override : ); ;State TUpdateStatus

( : ; procedure varDeleteRecord Retry boolean

var override : ); ;State TUpdateStatus

( : ;procedure varModifyRecord Retry boolean

: ); ;var overrideState TUpdateStatus

// procedure UnmodifiedRecord(var Retry:boolean;

var State:TUpdateStatus); override;

end;

mt AppendRecord. ();

mt StartTransaction. ;

mt Edit mt Post. . . ;

mt Delete. ;

mt Rollback. ;


mt StartTransaction. ; // First level of transaction


mt StartTransaction. ; // Second level of transaction

mt Delete. ;

mt Rollback. ; // Back to the edited record again

mt Rollback. ; // Back to the original appended record


mt StartTransaction. ; // First level of transaction


mt StartTransaction. ; // Second level of transaction

mt Delete. ;

mt Commit. ; // We will keep the delete but not the

edited record.

mt Rollback. ; // Ah no.. we will revert to the original

appended record

mt CheckPoint. ;

mt Undo. ;


COMPONENTS

DEVELOPERS4

procedure var var . ( : ; : );TMyDeltaHandler InsertRecord Retry boolean State TUpdateStatus

var string : ; , : ; : ;i integer s sv v variant

begin

:= ;s ''

:= - for to doi FieldCount0 1

begin

:= [ ];v Values i

( ( )) :=if then VarIsNull v sv ''

( [ ]. ) :=else if not in then Fields i DataType kbmBinaryTypes sv v

else

:= ;sv ''

:= + + ;s s sv ' '

;end

( ( ,[ ]));ShowMessage Format s'Inserted record (%s)'

end;

procedure var var . ( : ; : );TMyDeltaHandler DeleteRecord Retry boolean State TUpdateStatus

var string : ; , : ; : ;i integer s sv v variant

begin

:= ;s ''


begin

:= [ ];v Values i



:= ;else sv ''

:= + + ;s s sv ' '

;end

( ( ,[ ]));ShowMessage Format s'Deleted record (%s)'

end;

procedure var var . ( : ; : );TMyDeltaHandler ModifyRecord Retry boolean State TUpdateStatus

var stringi integer s1 s2 sv v variant: ; , , : ; : ;

begin

:= ;s1 ''

:= ;s2 ''


begin

:= [ ];v Values i



:= ;else sv ''

:= + + ;s1 s1 sv ' '

:= [ ];v OrigValues i



:= ;else sv ''

:= + + ;s2 s2 sv ' '

;end

( ( ,[ , ]));ShowMessage Format s2 s1'Modified record (%s) to (%s)'

end;

//procedure TMyDeltaHandler.UnmodifiedRecord(var Retry:boolean; var State:TUpdateStatus);

//begin

//end;

And :you start the deltahandler like this

var

: ;myDeltaHandler TMyDeltaHandler

begin

:= . ( );myDeltaHandler TMyDeltaHandler Create nil

try

. := ;mt DeltaHandler myDeltaHandler

. ;mt Resolve

finally

. := ;mt DeltaHandler nil

. ;myDeltaHandler Free

;end

end;



DEVELOPERS444

LOADING DATA

Data can be loaded into a memory table in multiple

ways:

1. Inserting them one record at a time

2. Loading the data from another dataset

3. Loading the data from a file

As for 1, you can use the ordinary Insert/Post, or

AppendRecord methods.

As for 2, you can use:

mt.LoadFromDataset(anotherdataset,

[mtcpoStructure]);

This will make your memory table have the same

fields as the anotherdataset (due to the

mtcpoStructure option), and load a copy of all data

from the dataset into the memory table.

A number of copy options exists:

mtcpoStructure:

Copies table structure (field definitions) from the

source. Any field definitions you had in your local

memory table are removed.

mtcpoOnlyActiveFields:

Only field definitions in the source that actually

are represented as a true field, are copied.

mtcpoProperties:

Copy over field properties like DisplayWidth,

DisplayLabel, Required, ReadOnly,

Visible, DefaultExpression,

Alignment, ProviderFlags, Lookup,

LookupCache, LookupDataset,

LookupKeyFields, LookupResultField,

KeyFields, DisplayFormat,

EditFormat, MaxValue, MinValue,

DisplayValues, TransLiterate,

Precision, Currency and BlobType.

mtcpoLookup:

Also copy lookup fields from the source dataset.

mtcpoCalculated:

Also copy calculated fields from the source

dataset.

mtcpoAppend:

Append records to the existing records in the

memory table. This cant be used with

mtcpoStructure or mtcpoProperties.

mtcpoFieldIndex: Copy the index position of

fields to ensure field order is the same as in the

original.

mtcpoDontDisableIndexes:

Default a batch insertion is made, where indexes

are only updated after all records have been

copied over. However if you for example have an

index with unique constraint on a field, then you

might want to have your copy to stop early, if

there is a duplicate of that field value.

That require that the indexes are updated on the

fly for each record. Standard Edition will have a

larger performance penalty than Professional

Edition of kbmMemTable for this situation.

mtcpoIgnoreErrors:

Ignore any copy errors instead of stopping

copying.

mtcpoLookupAsData:

Copy over lookup fields a regular datafield, with

its matching data.

mtcpoCalculatedAsData:

Copy over calculated fields as a regular data field

with its calculated data.

mtcpoStringAsWideString:

Assume that source ftString, ftFixedChar fields

should be created as ftWideString in your local

memory table.

mtcpoWideStringUTF8:

If string/character/memo fields in the source

dataset do not match wideness of the same in

the destination dataset (your local memory table),

then automatically encode or decode to/from

UTF8. Eg.

If the source dataset has field 'str1' which is a

ftWideString field, while the local

memorytable has the same field 'str1' defined as a

ftString field, it will UTF8 encode the data

coming from the source field before putting it into

the destination field. Similarly if a source field is

of type ftString, but the destination is of type

ftWideString then an UTF8 decoding will take

place before putting the value into the destination

field. The same takes place with

ftMemo/ftWideMemo and

ftFixedChar/ftWideFixedChar.

If you already have an existing field structure in

your local memory table, and want to copy fields

from a source dataset, where the field names do

not match, you can take advantage of field name

mapping.

Eg.

This is telling kbmMemTable that the local field

named str1 is called str_1 in the source table etc.

As for option 3, loading from a file, you can use:

mt.LoadFromFileViaFormat('somefilename',

someformatinstance);

mt LoadFromDataset anotherdataset. ( ,

[], );'str1=str_1;int2=int_2'


Issue Nr 6 2015 BLAISE PASCAL MAGAZINE45COMPONENTS

DEVELOPERS4

Its practical if the data of field fld1 is required to

be quickly accessible in your string list. In this

example, s will only contain the value of fld2. If

you need the field value both as an object and as

part of the string, you can include it twice in your

field list.

Eg. 'fld1;fld1;fld2'.

A different extraction method is GetRows which

returns requested fields as an array of an array of

variant.

This will return an array of variant containing an

array of variant of size 1, with the contents of the

field with FieldNo=1.

Eg. v[0,0] is the value of the first field of the first

record.

Since we specified to ask for kbmGetRowsRest

number of records, starting with

kbmBookMarkFirst, then we are essentially

asking for all records. If we wanted to start from

another place in the record set, you should

provide a TBookmark value for that place.

That is created by navigating to the record, then

use the Bookmark function to obtain a bookmark

for exactly that place in the record set.

And by providing a number instead of

kbmGetRowsRest, you can limit the number of

records returned to that particular count.

The last argument of GetRows, can either be a

field number, a field name or an array of field

numbers or an array of field names.

someformatinstance is an instance of a stream

format component. kbmMemTable comes with

two, TkbmMemBinaryStreamFormat and

TkbmMemCSVStreamFormat.

The binary stream format class stores data

compactly and efficiently, and is giving the

fastest performance. However the CSV stream

format class allows you to read in (and write out)

comma separated formatted data, for easy

integration with other external systems.

kbmMW (our middleware product) comes with

additional stream formatters for XML and JSON.

Each stream format instance have a number of

flags that can be set, to tell how it's supposed to

handle the reading or writing of data, but that is

a story for another time.

sf TkbmMemCSVStreamFormat Create:= . ( );nil

try

. ( , )mt LoadFromFileViaFormat sf'.\mydata.csv'

finally

. ;sf Free

end;

EXTRACTING DATA

A number of methods exist for easily extract data

from a memory table.

mt.Extract('fld1;fld2',slData);

This one will extract the fields fld1 and fld2 for all

records to a TStrings instance (TStringList

typically) that you will have to create beforehand.

If the optional AFormat string is given, each line

in TStrings will be formatted according to that

format. If none is given, then all field values for a

record will be separated by a space.

The above example will thus put a space between

the value of fld1 and fld2 for each record/line.

mt.Extract('fld1;fld2',slData,'%s=%s');

This will put an equal

sign between the value

of fld1 and fld2.

It is also possible to specify that the first field (in

the following case fld1) should be stored as an

object for the text line in the strings list.

mt.Extract('fld1;fld2',slData,'',true);

Then you can access that value by:var v variant: ;

begin

:= ( . [ ] v TkbmVariantObject slData Objects i

). ;Value

:= . [ ];s slData Strings i

end;

var v variant: ;

begin

:= . ( , , );v mt GetRows kbmGetRowsRest kbmBookMarkFirst 1

end;

v mt GetRows kbmGetRowsRest kbmBookMarkFirst VarArrayOf:= . ( , , ( , ));'fld1' 'fld2'



DEVELOPERS446

STATISTICS

kbmMemTable contains a number of

functions to allow very fast grouping and

calculation of statistical values for your

data. Returning a sum of fld1 for all records

visible in the current index.

v will contain the sum of fld1 for all records in the

current index.

v[0] will contain the sum of fld1 for all records in

the current index. v[1] will contain the standard

deviation for the records, and v[2] will contain the

count of records. The following functions can be

used:

MAX, MIN, AVG, COUNT, SUM, STDDEV

and USR1,USR2,USR3.

The USRx f

blaisepascalmagazine 44 uk

Documents

n d r o i d

e s t u d i o

n d o w s

g e s f o r

i n u xoverview of delphi

o x y g e n e

i n u xissue nr

delphi history page