Paul Sanderson wrote: > Whilst building a new app I created an index on every column some of which > were empty.
And with "empty", you mean that every value in that column is NULL? > The database is reasonably large (400K rows) and I notcied that > it seems to take as long to create an index on a column in which all the > rows are empty as it does on one in which all the rows are unique. SQLite still has to read all records to determine what the values are, and to create the index with all these values and RowIDs. That these values all happen to have the same value does not change the algorithm. Regards, Clemens _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
