What to consider with dirty reads?

ta5

Hello
What has to be considered because of dirty reads in "normal" C/AL development?
What has to be considered in special scenario, for example with NAS (see also here viewtopic.php?f=23&t=44676)

Any comments and/or code snippets would be appreciated!

Thanks in advance
Thomas

Slawek_Guzek

Do you have any specific problem to solve?

I cannot see any single rule of thumb - do this and that and you will be fine. Each particular problem needs specific solution.

Example: client 1 generates looooong journal (say generates year closure per 3 dimensions on 5m G/L Entry table).

Lines generated by client 1 are visible in other sessions, even if process at client 1 has not been finished and data hasn't been commited.

This is good behaviour if you want to just 'see' what is going on from another client session. No problems here. The same behaviour is bad if you have some NAS routine exporting or checking general journal table, as this routine will have access to uncommitted data. The same thing may or may not be considered a problem. The same problem may be addressed in different ways.

Kind regards,
Slawek

ta5

Hm, good question.
No particular problem at the moment, the reason for my question is this: Somebody brought the topic of dirty reads to my attention. I just wanted to find out what to consider in everydays programming task, of course without getting paranoid
Any tips and tricks will be appreciated.
Thanks
Thomas