>>>A funny index discovery I've made:
>>>
>>>index on something tag A
>>>index on something else tag A
>>>
>>>After the second index, the correct records show in browse. But when I run through the records with Scan, For, While, the wrong number of records is sometimes processed, particularly if there are non-unique keys after the second index. But it works fine with:
>>>
>>>index on something tag A
>>>Delete tag A
>>>index on something else tag A
>>>
>>>Designed behavior, or ??
>>
>>If it's really 'designed behaviour', it wouldn't be really 'little', because Reindex works this way. Maybe it's your_table_oriented?
>
>But maybe Reindex has the Delete Tag built in to it?

I don't think so, it would make costless 'UT-propagated' idea that DELETE TAG and then indexing is much better fix then REINDEX in a case of data corruption.