Denis Samson wrote: > I sincerely hope either or both versions work, because > I really need to bulkinsert tens of thousands of rows > into my database!
As a immediate workaround (or even permanent solution?) have you considered writing your data to a delimited file, and using BCP (or MySQL's equivalent) to load it into the DB? BCP et al are designed specifically to bulk load data efficiently (as the DB + driver permits, anyway) Concretely, I'm suggesting something along these lines: load'task csv' DATA writecsv fn =. jpath '~temp/bulkins.csv' fork 'bcp "', fn ,'"' 2!:55 <fn NB. Cleanup file. CAREFUL! Of course, you may have to play with the parameters to BCP or choose another delimiter from csv (pipes | are popular). If writing a temporary file is anathema (I can understand that), and your bcp supports it, you could try passing the delimited data to its stdin (using dyadic spawn in place of monadic fork), so it's all kept in memory. -Dan ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
