> Importing data and blobs via TIB_Query

Why? Why not use TIB_Cursor or TIB_DSQL?

There is an example of blobinserts in the sample apps.

> Now to speed things up, I use CachedUpdates := True
> with ApplyUpdates and Commit every 1000 records.

Why do you expect cached updates to speed things up? AFAIK all
updates are sent one network packet at a time (unlike selects which
the server packages up). So I really would not expect any benefit in
this situation.

> All data looks fine, but only every 1000th blob get
> imported, the first one in the cache.

I think I know the reason for this. TIB_Query is a buffered dataset
and not really designed for this sort of thing. I posted an
explanation a while ago about how TIB_Query (and Interbase) handles
new blobs - I can try to dig it out if you are interested.

However I do wonder whether this reveals a "bug" in the cached update
processing when using blobs. Jason?


Geoff Worboys
Telesis Computing