The problem manifests in at least TWO situations (at least on my setup):
1. When using the SEARCH function (button with 'binoculars' icon)
- click on TABLE to view
- click on the SEARCH icon and enter your search term in the 'Search for' box
- choose SINGLE FIELD in the 'Where to search' area
- in SETTINGS/Position choose 'Anywhere in the field' from the menu
- click on SEARCH until your term is found, and REPEAT this several times
- at some point (repeating 3 times?), the listing will show a particular found record REPEATED many times
- clicking SEARCH again just goes to the NEXT (REPEATED) record
- if I CLOSE the SEARCH WINDOW, and then I simply scroll down, the repeated records disappear off the top of the window (as expected)
- but then, if I scroll up the list, more multiple records appear, and it's a different record than the initial repeating record
2. If one simply OPENS a TABLE as a table listing, the problem also manifests:
- click on a database TABLE in Base to open the listing
- scroll down until you reach the last of the first batch of records (99 records)
- then scroll up, and you might see these 'spurious' repeating entries
It looks like there is something seriously wrong with display mechanism for database/BASE tables in LO. I opened the same database in OpenOffice 3.3, and there are NO problems at all! I'm using a Mac with LO 3.5.3. The problem does NOT seem to affect LO 3.4.6 on my setup. (By the way, I'm running the database showing this problem in 'file mode' but I suspect the other users reporting this problem are using the standard "Embedded mode' configuration.) It looks like it's a screen 'rendering' problem and, if ubiquitous, would pretty much cripple any TABLE editing though the listing.
I'd be interested to hear if anyone else can replicate this problem or if there are any other USER or BUG reports about it.
Re: Repeating Display of BASE Table entries in LO 3.5
The TABLE GRID display problem reported above is still present in LO 3.5.4 (Mac OS). Can anyone else confirm this problem? There was a bug report about it a few weeks ago, but it seems to have 'disappeared'. I reckon it's a fairly serious problem (for BASE users) and should be fixed asap. It's a LO 3.5.x show-stopper for me.
Yes, that is what one can infer from the bug report. We shall have to
wait and see whether it causes a performance hit, especially with
Java-based db connections, including the built-in hsqldb.
Forcing refresh of the fetched rows will necessarily reduce performance,
but the original solution which caused the display problem in the first
place, if I understand the corresponding OOo bug report correctly, was
an incomplete hack anyway, because both OOo and LibO load all of the
dataset into memory, even though only what is shown in the data window