#1 2010-05-30 15:32:01
- gerdlorscheid
- New member
- Registered: 2010-05-30
- Posts: 2
- Profile
CA11 missing basic database functionality
I tried to copy the content of a PGN database into a CA database. First a nice progress window is opened in the middle of my screen. I cannot move or colaps it. At some point the import stops. Some investigation showed that there was an error popup window behind the progress window, so it was not accessible. Again one trick (fast click on the program icon in the status bar shows this windows for less than a second) allowed to read that in one of the 1.500.000 games but not in which one the brackets are not ok. May happen even with a chessbase export. The next trick allowed to aknowledge this message. Now the import database was in an inconsistent state. It was not possible to open or repair this database.
It may happen that some of the games to import are corrupt or that CA is not able to parse a certain game. But essential functionality then is:
- CA continues import without this game and notifies which game could not be imported for which reason
or
- CI stops before this game and shows a useful error message
and in both cases
- The CA database is in a consistent state afterwards.
Calling a programm not fullfilling these requirements a database program is just ridicolous.
By the way: SCID is importing PGN into their native format much quicker then CA. It is open source, so CA developers should have a look...
Offline
#2 2010-05-31 12:56:22
Re: CA11 missing basic database functionality
gerdlorscheid wrote:
I tried to copy the content of a PGN database into a CA database. First a nice progress window is opened in the middle of my screen. I cannot move or colaps it. At some point the import stops. Some investigation showed that there was an error popup window behind the progress window, so it was not accessible. Again one trick (fast click on the program icon in the status bar shows this windows for less than a second) allowed to read that in one of the 1.500.000 games but not in which one the brackets are not ok. May happen even with a chessbase export. The next trick allowed to aknowledge this message. Now the import database was in an inconsistent state. It was not possible to open or repair this database.
It may happen that some of the games to import are corrupt or that CA is not able to parse a certain game. But essential functionality then is:
- CA continues import without this game and notifies which game could not be imported for which reason
or
- CI stops before this game and shows a useful error message
and in both cases
- The CA database is in a consistent state afterwards.
Calling a programm not fullfilling these requirements a database program is just ridicolous.
By the way: SCID is importing PGN into their native format much quicker then CA. It is open source, so CA developers should have a look...
==============
I will say about your suggestion to our programer
In current time ca indeed can't find incorrect game in PGN file and continue converting after automatic deleting such game
Also I think that better to check PGN database on such error games before convertation, if you don't sure that your database have no incorrect games.
If ca will check every game during converting, process may be too long.
Offline
#3 2010-05-31 14:59:05
- gerdlorscheid
- New member
- Registered: 2010-05-30
- Posts: 2
- Profile
Re: CA11 missing basic database functionality
I wrote already that I do not care whether the import stops or continues without this game. So I do not understand your point.
I mentioned that I get no hint which PGN game was corrupt and that the CA database is corrupt afterwards and cannot be repaired.
The PGN import is so slow. it would not matter if it becomes a bit slower ;-) As I told before, SCID does implement all these requirements. It does import the whole database and pops up a list of the inconsistencies afterwards. And it is much faster than CA. Please have a look.
Some programming hints: It is not necessary to check each PGN-Game first for consistency. What is required to implement is that the conversion first is done in memory and then - if no error comes back - written in one transaction to the database in a way that the repair tool is able to recover write errors.
Offline