On 22 Oct 2012, at 8:35pm, Paul van Helden <p...@planetgis.co.za> wrote:
>> It would be possible to implement TRUNCATE TABLE on top of that, but >> this would be only syntactic sugar. > > ..or better portability. TRUNCATE TABLE works (since only a few years) > nearly everywhere. So when writing portable applications it seems a bit > silly to make an exception for SQLite if the solution is that simple and > won't break anything. The problem is that TRUNCATE TABLE means different things in different SQL implementations. Yes, they do all delete all the rows of a table, but some will do it even in violation of foreign key requirements and some won't; some allow it to be an element of a transaction whereas some consider it to be a schema-changing operation which can violate transaction structure; some obey all TRIGGERs involved in deleting rows, others just execute it ignoring TRIGGERs. On 22 Oct 2012, at 9:44pm, Guillaume Saumure <gsaumur...@videotron.ca> wrote: > Personally, I have ask the question why TRUNCATE TABLE don't work because > this command is not listed on http://www.sqlite.org/omitted.html Because TRUNCATE isn't in SQL92, the version of SQL that this page starts from, as clearly stated on that page. It was introduced as a standard only in 2008, and since, as I wrote above, different engines do different things about it, anyone writing for portability across different SQL engines never uses it because they can never be sure what it does. Simon. _______________________________________________ sqlite-users mailing list sqlite-users@sqlite.org http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
