I can't say for certain, but there seems to be some sort of limit on the number of items that can be in a value list.

Say I have a list which is based upon a field:

[car parts::serialNo]

And then I use

[color:red]ValueListItems ( fileName; valuelist )

There will sometimes be items missing.

If instead I use

[color:red]List ( car parts::serialNo )

This problem does not occur. (Before anyone asks, the reason for not using list in the first place is for backwards-compatibility).

Has anyone seen anything similar? Is there a known limit on value list items?

I am not sure what you mean by missing. A value list shows only unique items, and this is reflected in the result of ValueListItems() too, while the result of List() will include duplicates.

By missing, I mean if you do a PatternCount(Query)>0 on both datasets it will sometimes evaluate as false on the ValueList type, but true on the List type.

Hope that makes sense

That could easily happen if the value list is defined to show only a second field, or to sort by a second field. Again, items with duplicate values in the second field would be removed from the list. If that's not the case, post a file showing the problem.

EDIT:

Another possible cause is values longer than the index limit (110 characters, IIRC). Values that are identical up to the limit are considered duplicates.

Edited April 23, 200916 yr by Guest

It could also have to do with the context in which List() is run.

ValueListItems() pulls each unique value from the entire index, List() grabs all the values from the set of records related to the current record.

Edited April 23, 200916 yr by Guest

That could easily happen if the value list is defined to show only a second field, or to sort by a second field. Again, items with duplicate values in the second field would be removed from the list. If that's not the case, post a file showing the problem.

Neither of those things are the case here. Unfortunately I cannot post the file for IP reasons. I will try to find a way to replicate the problem in a clean file.

Another possible cause is values longer than the index limit (110 characters, IIRC). Values that are identical up to the limit are considered duplicates.

This sounds plausible. I will check what the average length of the values is.

Hmm.

Well in the process of creating a clean file, I've been completely unable to replicate the problem thus far.

The only oddity I could find was that the indexing on the field used in the value list was set to "minimal" which I suspect might account for the problem somehow.