
Timeout and imcomplete PRO ISAM data update
Started by Surajit, Jun 10 2004 09:34 AM
21 replies to this topic
#16
Posted 15 June 2004 - 05:11 PM
Bill,
Thanks for the info.
2 follow ups:
1) Did you have to change the settings in /etc/isamdef at all?
2) Are you at all concerned by how much data you put in your memory files? For example, a recalc routine that currently uses a ProISAM temp file may be storing upwards of 50,000 rows for a very large client. Is this a concern?
Thanks,
Joseph
Thanks for the info.
2 follow ups:
1) Did you have to change the settings in /etc/isamdef at all?
2) Are you at all concerned by how much data you put in your memory files? For example, a recalc routine that currently uses a ProISAM temp file may be storing upwards of 50,000 rows for a very large client. Is this a concern?
Thanks,
Joseph
#17
Posted 15 June 2004 - 06:10 PM

As for question 2, we have employed a concept foriegn to most proiv shops, data normalization with foriegn keys. So 50,000 rows of normalized data is not much data. If only store the keys to your data then accessing the data for recalculate is easy.
HTH Bill.
#18
Posted 15 June 2004 - 06:37 PM
Bill,
Thanks for the info.
...
Data normalization - is that where you give the same column name for the client name in each of the six tables it's stored in - OR - where you only let users input normal client names like Bob and Sue and not Dweezle or Moonbeam?
Regards,
Joseph
Thanks for the info.
...
Data normalization - is that where you give the same column name for the client name in each of the six tables it's stored in - OR - where you only let users input normal client names like Bob and Sue and not Dweezle or Moonbeam?

Regards,
Joseph
#19
Posted 15 June 2004 - 07:03 PM

Now you can go crazy and got to the twelth level of normilazition which may cause a twelve to fifteen table join, as I have seen, but I do not recomend this. This was done by a RDMS instructor that I had and he was a nut case but RDMS smart.
HTH
Bill
#20
Posted 15 June 2004 - 11:08 PM
Bill,
Right, normalizing - next you're going to say that you use column names that actually describe the data that is being stored in them!
I was just offering some sarcasm.
I guess the wink icon was not a big enough tell... Next time, I'll try to use a better graphic...
Regards,
Joseph
Right, normalizing - next you're going to say that you use column names that actually describe the data that is being stored in them!

I was just offering some sarcasm.



Regards,
Joseph
#21
Posted 18 June 2004 - 08:28 AM
How do you release the space??
Setting the clear flag or deleting the records does not release the memory used...
Depends on what you mean by release. The clear flag seems to make the memory available for other MKY files in your session, although, as you say, there seems to be a bug and it should release it back to the o/s.
So, having spent more time testing this and although I've not had any
related user problems, I'm going to only use MKY for predictable small
data sets until this is sorted out.
Edited by Chris Mackenzie, 18 June 2004 - 08:32 AM.
The content and views expressed in this message are those
of the poster and do not represent those of any organisation.
of the poster and do not represent those of any organisation.
#22
Posted 18 June 2004 - 08:37 AM
Yep, I meant back to the O/S.
I only use them for small amounts of records, and use them very carefully.
If you get the chance, I would suggest you contract ProIV LTD about the releasing memory problem, because think the response I got from them was that 'thats how it works'.
To make Memory files more practical, I think this needs to be sorted.... and it cant be that difficult to sort out surely.
Rob D.
I only use them for small amounts of records, and use them very carefully.
If you get the chance, I would suggest you contract ProIV LTD about the releasing memory problem, because think the response I got from them was that 'thats how it works'.
To make Memory files more practical, I think this needs to be sorted.... and it cant be that difficult to sort out surely.
Rob D.
Reply to this topic

0 user(s) are reading this topic
0 members, 0 guests, 0 anonymous users