I have a file with a corrupt index. When selecting data using the indexed field it returns no data. When attaching the file in arev and doing the same thing I get a message saying "Link list in table

!TABLENAME is broken and transactions cannot be properly flushed"

I want to know how to check for this kind of error in OI.

GET_STATUS after using RLIST(criteria,5,"","","") returns no error.

GET_STATUS after UPDATE_INDEX(…… also returns no error but the index is still corrupt.

If Arev can tell me that the index is corrupt why can't OI?

I can fix the problem by removing the indexes, manually deleting the !TABLENAME file and re-creating the indexes. (Rebuild doesn't fix it)

I just want to know how to detect the problem in OI.

I know, I don't want much do I?

Paul,

I'd like to know too. There are a couple of indexes in my new app that could cause serious problems if they failed. (Like tying me up with a lot of user support.)

- Oystein -

I have a routine in AREV that diagnoses and fixes broken link lists in the !filename. I just haven't gotten around to re-writing it for OI. If some eager beaver is interested in having a crack at it, send me your e-mail address and I'll zip up the source code for it. It may need to have some surgery to make it work in a Runtine OI because it tries to create dict entries in the VOC of the account. It also creates a popup on the fly to allow for selecting tables to analyze and fix. Give me a shout if you're interested.

Don Miller

C3 Inc.

Don-

Is this the utility from Mike Pope in 92?

Mike

Don,

I'm interested.

In one way I'm more concerned with detecting damage than reparing it. E.g I have a field that should have a unique value, if certain problems are to be avoided. To ensure uniqueness I check its index before allowing a new value.

In principle I'd like to know all things that can go wrong with an index table, and all the symptoms. For all I know broken link lists is just one of many ailments an index table can suffer from.

I'm curious to know if it takes a lot of effort (time) to discover if there's something wrong with an index file. That will decide where I should implement the check - each time a row is written or at some longer interval.

My email address is .

- Oystein -

Don't think so. Maybe pieces .. but Mike's stopped working at AREV 2.x and mine goes through AREV 3.X. Did it so long ago that I don't remember where it all came from.

Don

Oystein ..

There are many subtle things that can go wrong in indexes. For me, the greatest bugaboo is in numeric data (including leading zero's) in b-trees or cross-references (which is just another form of b-tree). Broken link-lists in the !filename are pretty easy. Traversing an index forward or backwards to check for holes, etc. is not simple, but can be done. It's very difficult to spot errors in an index unless you're dealing with a file key or a data-field that is required in all rows (null variables not allowed). Then you can check the number of entries in the b-tree against the file-count by using Get_Reccount or whatever to get a file count. A mismatch indicates either a damaged index or a corrupt file header on the data file itself. Anyway, I'll e-mail you the source code for FIXLIST. It's in REV format so I'll zip it up.

Good luck,

Don Miller

see thru ink

But some people enjoy being tied up :

With user support? Are you sure? (Do they have web sites for this kind of stuff?)

(Or did I say something ungrammatical or invent my own idiom or something?)

View this thread on the forum...