Recent content by Petr Danes

I haven't dug into it enough to isolate a specific set of circumstances. But I don't understand your statement here, "...all records on a page are overwritten with the first record." I thought the space was just marked as available. Why would a 'first record' be written everywhere? That seems...

Thank you, but no - it is not a timing problem. This erroneous value persists even after closing and reopening the database. Only a C&R fixes it.

I don't blame you for only skimming this thread - it got somewhat bogged down by irrelevant BS at one point. However, you missed what was actually happening. The Delete command does NOT fail. It works fine. What fails is the subsequent value returned by the syntax I used. The table is completely...

Crap - spoke too soon. A few more hours of work and the problem is back. In the same table, with the same incorrect ghost count, and a new table added to the mix, also with a non-zero count on an empty table. Whatever is going on, it's not reliable enough for me to use. I'm giving up for now -...

One more note to this. After replying to Colin about Access obviously using two different methods, I got to wondering about performance, so I wrote a tiny test routine to count the records in the largest table in this app, almost 200,000 records. Here it is. I ran it three ways, each time...

True, I don't really have an explanation - corruption is such a vague and all-encompassing word for so many problems, and similar-appearing problems can have vastly different causes. But trying to go through the machine code of exactly what was happening is more than I am willing to tackle...

If you want to make a difference, pick up where Stephen Lebans left off http://www.lebans.com/. That guy knew his way around the internals of Access like nobody else. God only knows where he got the information to build the stuff he did, most of which remains functional to this day. I've used...

Well, it looks like that was it - a corrupt table definition. I removed the relationship, deleted the table, did a C&R, decompile, another C&R, created a new table with the same definition and added the relationship back in, and did a final C&R. Loaded the tables in the standard manner, using...

Thank you. Yes, using "*" might be more readable. I've just gotten into the habit of using "1". I'd like to use an empty string, to indicate that I have no interest in any field, but that throws an error. As for the original topic, I have no clue why it doesn't work, and apparently nobody else...

No. All machines on which I run this act the same way. The end user is stuck in 2007, so when I develop a new version on my modern laptop with 365, I run it through an old server on which I also still have 2007, and then send it to him. The problem occurs on all of them. The only thing I can...

It is NOT caused by any non-standard procedures, and I am not using any non-standard procedures anyway. I am DELETING ALL RECORDS FROM A TABLE. Period. There is nothing non-standard about an SQL Delete command. What else is going on around that is completely irrelevant to the topic. Yes, I do...

Jesus - nobody is asking you to. But being pedantic about anything is usually a mistake. There are situations where a monolith is a better solution. If you insist on avoiding it for no reason except that you think splitting is 'always the right way', you add extra work for yourself and extra...

Yeah, I figured someone would bring this up. Splitting is one way of doing things, but not the only way, and it is certainly not 'the problem' here. I have intentionally not split this one, because it needs to remain portable, and the single-file design is far better for that. And I am NOT...

See my answer in #38. I explain all that.

I have no idea what any of that means, but I'm getting the impression you're jerking me around for fun, and I'm not interested in that. Bye.