Filtered value list not showing all matches

[Kirsten]

I am working on a small (10 K records) database, and have encountered a problem that I hope is a programming oversight, and not data corruption. There are two files, with a many to many link created in a third. File one: people. File two: events. File three: join file.

In the events file, to add a new person to an event, I have created a value list made up of the letters of the alphabet. When I want to enter someone whose last name begins with 'a', I select 'a' from the value list, which drops it into a NameFilterGlobal field. I've made a relationship between the NameFilterGlobal and a calculated field in the People file that parses out the first letter of the last name. I also created a second value list, which is filtered on the relationship above. So, if I chose 'a' from the value list, then everyone whose last name begins with 'a' should show up in the value list (which I've attached to the "choose name" field in the Events file.)

This arrangement worked like a charm, and ensured that only a few hundred names would show up in the value list (based on the first letter of the last name chosen), instead of all 10,000 names in the People file. The problem is that recently, when setting the letter global to 'w', only half of the names show up in the filtered value list. Specifically, the "wa's".

In addition, sometimes one of the names will be filtered to be included in the value list, but the name won't actually display, only a blank line.

Any suggestions about this? I've checked the other letters of the alphabet, and there are other letters with more names (for example, 'c' has almost twice as many names as 'w'.) Thus I don't think this is a filtering maximum I've hit. And the names are distinct, so it's not an indexing issue.

Any suggestions are welcome.

Kirsten Masse

[Søren Dyhr]

With fm6 and below is the upper limit 64K for any values stored as text in a field or calc'field. Some extreemly crafty algoritms could however circumvent this limitation, by stuffing smaller chunks in repeating fields repetitions which all individually have this limit. I seem to remember that something is to be found in this forum showing this.

--sd

[Søren Dyhr]

I have reread you post and there might something else, is there somehow a way you could have modified the indexing method into ascii for one of the keys, your's is most likely to be US-English?:

--sd