On 14 Feb 2012, at 6:48pm, Marc L. Allen wrote: > Does the data type start at being the same as the column affinity in the > underlying table and then adjust if the actual data isn't of that type?
<http://www.sqlite.org/datatype3.html> section 2.3 gives examples of what happens. But an overriding rule is about losing data. So SQLite might try turning a string into an integer, then back to a string again. If it gets the same string back, then storing it as an integer didn't lose any data, so it'll do that. Bad example, but I hope it helps. > And, if I don't know the underlying table structure, I am guessing that I > should always treat INTEGER as a potential int64? Use sqlite3_column_int64() or sqlite3_column_int() from <http://www.sqlite.org/c3ref/column_blob.html> whichever suits you, yielding different sized integers. Simon. _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users