I found it interesting. Thanks Eric. >>(reconstructed from archives i accidentally deleted the copy in my mailbox >>>Daniel Clark writes: >>>I don't see how LIMIT would make a difference. LAST_INSERT_ID() only >>>returns one record. >>>But it's worth trying in a big loop to get timing numbers. >> >>Well, I decided to test this all out and see what happened. From what my >>little test was able to determine, there is no difference in speed from >>doing: >> >>SELECT LAST_INSERT_ID(); vs SELECT LAST_INSERT_ID() LIMIT 1; >> >>I used 100,000 selects as my benchmark number (which on my box here took >>about 40 seconds per run to complete). Times were measured using unix time, >>and I averaged the user times and the difference was so small as to be >>negligable, then i decided to measure the same thing but with >>SELECT LAST_INSERT_ID() FROM table; and it was approximately 25% slower (but >>im assuming as the table grew in length so would the gap in speed. >> >>Conclusion: it makes no difference positive or negative including the LIMIT, >>but as Michael pointed out including the FROM clause causes a big penalty. >>I hope someone else finds this slightly interesting. For completeness im >>including the program i wrote to test this. Your mileage may vary, it needs >>a little configuring for your particular setup before you can run tests. >>(database name, user/password etc) you have to manually change the one line >>inside the loop to call whichever function you want to test. And its also >>handy to change the one print statement before the loop. >> >>Eric
-- MySQL General Mailing List For list archives: http://lists.mysql.com/mysql To unsubscribe: http://lists.mysql.com/[EMAIL PROTECTED]
