FiveTech Software tech support forums

Enrico Maria Giordano wrote:Can you send me the DBF and the CDX to reproduce the problem here, please?

EMG

SENT

Ok, thank you. As you can check for yourself, the CDX contains a tag with "id" key. So it's just a corrupted CDX. Please delete it, rebuild it and let me know.

EMG

I erase the cdx.
then I run the test made from James
it create the cdx
then open emagdbu , open customer database
Open the cdx and set the tag id
the same

The example from James is not valid for your customer.dbf because it builds the wrong tag:



Your customer.dbf doesn't have a field named "id".

EMG

yes of cource but...

why the code ( test.prg) compiled create the cdx if is not valid ?
we Know it's is not valid but the exe crete the cdx

I don't know, sorry.

EMG

Sorry if I sent the wrong customer file. It is disconcerting that the CDX was created without any error message. That would be hard to track down as we have just seen.

Silvio, were you able to get the sample to run now?

James

James,

I wanted understand why the command :



create the cdx and not make error

It seem we can index any field not exist into a dbf ...

I just realized that this syntax is uncorrect:



The correct one is:



Or, even better:



EMG

Enrico,

Thanks for finding that.

I did a search of all my source and I did find a couple of other times I used quotes. It would be nice if this generated an error. I must have been going nuts trying to figure out why those indexes weren't working.

Enrico, would this be a (x)Harbour issue. If so, would you mind reporting it?

James

This is not a bug. Is perfectly legal to build an index on a constant key:



EMG

Ok. I am curious of what use that would be?