Speed up find script

[jrRaid]

I have to change the find script in a 100 000+ record inhiretted file that performs slower and slower.

I found out that the script is searching on an unstored calculated related field.

What or how is the best way to speed this up?

Whatis the way to make that unstored field somewhat 'stored' at the search moment?

TIA

[Matthew F]

Did you try turning on indexing for the calculated field?

[jrRaid]

Then the 'unstored' feature is lost.....

And that is impossible because it reference a related field....

Edited June 22, 2010 by Guest

[Lee Smith]

Hi jrRaid,

That's how relationships work. The right side, or Related record side (this is were the actual data resides that you are viewing in the Portal), must be stored.

Understanding how relationships work is a big part of developing files beyond a simple rolodex, and the more you understand the basics, the easier it will become in making your files function as you envision they should.

Don't forget that you can upload a demo, or copy, of your files, which will help us understand what you are trying to do, and then address our answers to fit your files .

Lee

[Matthew F]

OK. I wasn't sure if you meant a related calculated field, or a calculation using a related field.

It seems like the answer then would be to keep all of the data for the calculation on the current (not the related) table.

If this is really not practical then perhaps you could perform sequential searches. First perform a search using local (indexed) data. Then further constrain the found set using additional data in the related table.

[jrRaid]

It is clear that I look at a design flaw here.

I don't want to rewrite the whole application.

This is a real search for a workaround.

[Lee Smith]

As I suggested before, upload a copy of your files.

If you are unsure of how to do this, take a look at this link for more information. Link

[Matthew F]

If you take Lee's advice and upload a copy of your files you might want to first make a copy that strips out aspects that are irrelevant to your question. You can replace the real data with fabricated data.

In stripping down a database, sometimes the flaws in the design become readily apparent and you might end up answering your own question.