[jan marco alkema]

Quote:

Well … do you have an index on `lastname` column? If not the server will need to do a full table scan.

I have index.

The problem is to make a selection between first distinct ‘lastname’ and the last distinct ‘lastname’ in the filter window. When there are a lot of distinct “last names”.

I imported the distinct last names in Office2007 and OpenOffice as much as possible. Both filters don’t work properly. For example: Office2007 listed 22 last names in the filter only. See included files for an impression.

OpenOffice has a “Standard Filter”. You can filter on ranges, but I find it not very user friendly. Maybe I have tried not enough…

[jan marco alkema]

Quote:

Yeah .. and one good method is to write a query with a WHERE clause ..

I know how to do it, only a program ‘to call some else’, should be easy to use. My mother does not know what a “where clause” is.

Number of Filter cases:

Quote:

USE foondump2005;SELECT COUNT(DISTINCT lastname) FROM white_subscriber;

1,297,342

Quote:

I also did not tell that we will never do it – only we will add more single column filtering options soon.

I am very curious, how “you” make a fast filter on 1,297,342 last names (= 1 column).

[jan marco alkema]

Quote:

There are other ways (check Office programs like MS Office and Open Office)

Thanks for your feedback. Filter technique in MS Office (Excel) is a good one.

N.B. If you have > 5,000,000 telephone records, it is very usefull to have a good filter technique to find a person.

[jan marco alkema]

Quote:

We cannot not change a single bit in a BLOB (and not zip it either) – for the simple reason that it is the server that stores data and not the client.

It is quite normal that the client retrieves records from database server and the client could manipulate (insert, updata, delete) records (in the database server) also.

Quote:

If we manipulate what is stored it will fail with other clients (like web applications).

Databases engines are made for manipulation of data (records etc). If there are more than one client on a database, you can give some client less access to the database. Only read rights, etc. Or put an “application server” between users and database.

[jan marco alkema]

Hello DonQuichote,

I see a blob field as a name space http://en.wikipedia.org/wiki/Namespace

If you want to store files in it. You should put a structure around the files. There was a project ( http://www.codeproject.com/KB/files/nfsLibrary.aspx ) that ‘mounted’ a file system in 1 file = 1 blob field. Only the project has been removed. Maybe you can zip the files you want to store in a blob field. If you want to read it, unzip the blob field. A zip file can contain more than 1 files.

Greetings Jan Marco

[jan marco alkema]

Hoi Peter,

My record filter idea, I got from an old version of http://www.codeproject.com/KB/applications/todolist2.aspx

Make a window beneath the grid. Put automatically all fields of the table in the window, by putting “selection boxes” in the filter window. To filter the grid, type a value (for example field ‘Priority’ = 5) in a field of the selection box. If you do this, Only the records with priority 5 will be displayed. See included file as an example only.

Greeting Jan Marco

[jan marco alkema]

Hello Peter,

I have no request of changing the SQLyog program, only advise how to design with an other (business) program.

The telephone program is a good example to explain what my ideas are.

For example: You have a peer with 155 subscriber records. Nithin has a peer with 75 subscriber records and I have a peer with 90 subscriber records. How to do the database synchronization between the three peer?, without less publication of the subscriber records. Publication could encounter copy right violations.

Quote:

The problem from our point of view is that there will be a single user/customer for every detailed implementation of this. You want to start a call from a phone no. column – somebody else may want to open a browser from a URL column or open a HEX-editor from a BLOB column etc. etc.

If some one else or you change the value of a subscriber record. The old record will be placed in ‘subscriberarc’ table. Only a very short period there will be a problem with consistency, but there are semaphore and mutex routines to prevent this on 1 (virtual) PC.

Records have a ‘life cycle’, for instance a begin and end time. The VOIP call has a ‘life cycle’ (begin time stamp and end time stamp) also, you can determine which record is valid, the new record in subscriber table or the old one in ‘subscriberarc’ table. If a call begins a new record will be insert in a “voip” table. On this table the poll down menu will have relevant “options“, like ‘stop call’, ‘invite subscriber’, ‘resume call’, ‘transfer call’, etc. The integration with SQLyog is more intense than call an external program only.

Quote:

Also the 'content' is hard to identify in lots of cases. For instance I believe there is no international standard for phone numbers. How should SQLyog identify a string as a phone no? Is “99 1-4546-6789” a valid phone no. in Farawaystan?

The Phone numbers are used for searching the right entry in the table. You can see them as a “label” only. To call with the old “analog” telephone or to find a person. The peer IP is the key to communicate with other peers. If you start up the program (“SQLyog+wengo”) the public ip can be determinated very easily, if you are connected to internet. You search your own subscriber record in the subscriber table. In the pull down menu will be an option “set as default peer”. After that, your IP will be (automatically) synchronized with Nithin and my peer, by the synchronization of the subscriber table between the peers.

For database synchronization you should have the same table structure. I think of hashing the fields of a table. If you, Nithin and I have the same “table hash”, we could synchronize with each other.

Quote:

For your particular 'project' we do not know either if this specific program has an API that makes it possible to control it from other programs or if it takes command line parameters such that the phone no. can be passed as a parameter to the program call. That would also be an requirement of course.

Good one –)

Qutecom (Wengo) depends heavily on QT (I get it compiled, linked and executed with “qt-win-opensource-src-4.3.2” after a lot of study how to do it). I would get rid of QT as soon as possibly!

My strategy is to compile, link and test “qutecom-2-2-b816aa35e47cb”. This works at this moment. Now I am beginning to remove the (qutecom) dll’s, one after the other, by including the source in the project and remove the dll-dependency. After all dll’s are gone, I move the source (without QT sources) to the SQLyog source.

Quote:

I think in conclusion that you should not expect us to try to implement this. It would not be very hard to add an item to the GRID-context menu calling an external program. Anybody with reasonable C-programming skills can do it (based on the COMMUNITY code that is freely available) if it is the only thing that is required.

No, I think that integration should not be done by calling an extern program!

Conceptually an integration of SQLyog, Qutecom and Chrome (browser) would be nice.

1) SQLyog: The (table) interface to call, mail, order products, etc.

2) Qutecom: To communicate with other peers. “Some” peers are behind firewall. STUN concept, etc.

3) Chrome: Browsing the internet, very OS like solutions are programmed in Chrome. For Example: One program has different (sub) processes. The prio of threads are depending on the prio of the process (‘Real time’ VOIP calling to ‘batch’ database synchronization)

If these three programs are integrated, for example one good “worker thead pool concept” can be worked out, which is used by all ‘three’ program parts.

Greeting Jan Marco

[jan marco alkema]

>But I do not know if we can get reliable information from Windows itself.

Try example of GetProcessMemoryInfo: http://msdn.microsoft.com/en-us/library/aa366589(VS.85).aspx

>(what is no problem if you show 100,000's or rows (but could be an annoyance if you show only 20).

Make a tresh hold for remote host, if estimate size is higher (or lower) than x % of available memory, to do the check (or don't do) the check.

[jan marco alkema]

>When a query is sent the control is taken over by the MySQL API compiled into SQLyog and it does nothing else than fetching the rows. It simply assumes that the result set can be handled by the OS.

Conclusion: The application: SQLyog should manage the request!

I am not expert on this matter, but this are my suggestions, without al lot of study on this subject:

1) Make an estimation of the memory what is needed to show the table. You can get the average field length of a table in MySQL.

2) Compute the available amount of memory on the host. A ‘little bit’ old project for example http://www.jfitz.com/software/RAMpage/index could possibly be used. Maybe there are recent project with same functionality.

If (estimated memory < available memory) { Show records in Grid. } else { Initial: Select a restricted limit of records of the table. Depending on available memory and estimation of record length. Depending on entering control-Up/Down keys or WM_VSCROLL messages, select other parts of the large table. If I enter more control-Up/Down keys in a short period, the move of the display (shown Grid records) over the total records will go faster up or down wards, like a ‘train’. A train begins slow, but will go every time faster until the maximum speed has reached. }

[jan marco alkema]

Hello Navya,

I test http://www.webyog.com/downloads/SQLyog822Trial.exe on Vista en the latest svn release on XP.

I don't seen the bug any more. Thank you, for solving it –)

Jan Marco

[Author]

Posts