I have a search problem. When some of our users do a find for a certiain status("Revisions Needed" for example). The search returns all the records that have that status, but occasionally it either returns extra records or reports the correct records, but with statuses that show up as something other than the search criteria status.

My first thought was a cached data issue, but even after a restart of filemaker and the machine, they still get it. I can't replicate it, and we have tried a reinstall, including throwing out all the preferences on the machine. No luck.

Any ideas would be greatly appreciated.

~Mark

What is the data, what type of field. What is the search??

What is the data, what type of field. What is the search??

The field they are searching on is a text field with a drop down menu attached to it. The data is strict text no numbers, dates, etc.

The search is not fully scripted, there are two ways they can do it. they either enter find mode on the layout(for the ones who have figured that out) or there is a button that takes them to another layout where they fill in globals, click find, and it sets the relevant fields to the globals, and performs the find.

Either way they do it, the same thing happens.

~Mark

Hi mcyrulik

can you write here a search criteria (as an example) and some right and odd found set ?

Yes, it would be good to look at the file unless the file itself has gone corrupt, indexes are broken and needs to be recovered.

Hi mcyrulik

can you write here a search criteria (as an example) and some right and odd found set ?

The correct.pdf file shows the correct results. The incorrect.pdf file show 2 extra pages, 299 and 300. These are pages that are at different status, and have no reason to show up. These searches were done approximately 20 minutes apart, from the same machine using the same method.

The search criteria for both of these was the text "Revisions In Progress (OEC)" selected from a drop down menu.

Incorrect.pdf

Correct.pdf

If indexes were broken, or the file were corrupt, would that cause it to only happen to a few users? There are 3 machines this has happened on, I can't replicate it on my machine no matter how hard I try.

ok

I was able to look only at the Incorrect.pdf (as I have AR 6).

Can you say me how many records have the field with "1st Round Review" ?

(it seems as you searched for Rev )

7 are at 1st Round review.

ok

Can you do now manually a search in find mode with this: ==Revisions in Progress (OEC)

into the field STATUS

(it will be better if you make a new one in layout mode, without the popup menu)

I tried doing it your that way with the "==", and if they go in through the script, it searches this way anyway, and the results still show incorrectly. Even without the "==" if they are selecting a value from a dropdown list, would that still show up results that seem to make Revisions In progress (OEC) = 1st round review?

Edited September 21, 200520 yr by Guest

Hi,

Have you tried:

1)Change the two incorrect fields to another item of the popup menu, then commit records, then re-change them to "1st round review"

2)File-->File Maintenance...

3)Make a clone and import from the original ???

1)I have tried that in the past, but those statuses change pretty regularly. For example, I just got another PDF from a user and the search was done on the same status again, and this timepage 258 came up in 1st round review(it was pages 299 and 300) yesterday

2) no not yet, what exactly does the File maintenance do, I mena i know it optimizes, and compacts, but will either of those help my situation?

If indexes were broken, or the file were corrupt, would that cause it to only happen to a few users? There are 3 machines this has happened on, I can't replicate it on my machine no matter how hard I try.

This is why I was thinking about a simpler error...

but I had just re-read this post and you didn't specify if the machines share the same DB !

Do they ?

For the 2) point: yes it is true, but don't corrupts the file...from Help on Line:

Improvements will be especially noticeable when you are using large files and performing finds or sorts, or executing scripts that operate on all records

I'll wish that this do not need to you !

This is why I was thinking about a simpler error...

but I had just re-read this post and you didn't specify if the machines share the same DB !

Do they ?

Yes, it is a DB served off of an FM7 Advanced Server. All the clients and the server are updated to FM7v3.

And the incorrect found set is only in some client ?

yes, we have only found that it occurs on 3 machines so far.

Hi mcyrulik

I'm sorry but I can't go on with this topic, because I haven't any experience with FM Server.

I hope that someone else can give you some good point. ???

You mentioned global fields earlier. Some users MAY enter their search criteria into global fields for a scripted find. Are you sure these globals have not accidentially been pre-populated with some value? The value could be hidden if it were on line 2 of a field that was only showing 1 line.

Or could the records themselves be holding two or more values in the same method described above?

I thought of that, but what kind of throws that out is that when I get incorrect results it doesn't include ALL of the incorrect results. For example, in the above correct.pdf, it shows two pages at 1st round review. If you did a search for 1st round review, there were seven pages at that status. If there was data on a second line, or hidden wouldn't it show all seven?

Yep, I agree, I think it would.

Have you tried creating a new field and then copying the info from the first field over to the 2nd field? I wonder how the searches would come out on the 2nd field? It would be interesting to see what happens.