Chris Burke wrote: > Matthew Brand wrote: >> I have written a couple of emails about this JDB bug but was not too >> sure if it was a bug or just my lack of understanding. I think the >> output below shows very clearly that there is definitely a bug when >> trying to make a compound primary key whose columns reference another >> table. The code below results in duplicate values in a table which >> should only contain unique rows. >> >> The fact that at the end of the script the Reads__ command does not >> report row 'A';'B' as being in the table, but that further attempts to >> Insert__ 'A';'B' fail with duplicate key error might give a hint as to >> where the problem is. Insert__ might be doing the correct thing but it >> might be using Reads__ somewhere to find what is in the table already. >> It might be Reads__ that is not be working correctly. > > Thanks for the detailed report. I see the problem (with Reads, not > Insert) and hope to get a fix out tomorrow.
This bug is fixed now. I also made a related change - a scalar value for a varchar type is now raveled to a 1-element list. ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
