|Pathologically Eclectic Rubbish Lister|
The Naughty Nought "OEO" bluesby rinceWind (Monsignor)
|on Jan 20, 2005 at 15:40 UTC||Need Help??|
I've been bitten by a gotcha on return values from DBI execute (and Super search tells me that many others have been bitten by it in the past also :).
For a non-"SELECT" statement, "execute" returns the number of rows affected, if known. If no rows were affected, then "execute" returns ""0E0"", which Perl will treat as 0 but will regard as true. Note that it is not an error for no rows to be affected by a statement. If the number of rows affected is not known, then "execute" returns -1.
Consider the following code:
This is Sybase, and I want to run the query repeatedly adjusting tmpDate in temporary table #LAG until none of them fall on holidays for the particular country. But, this code loops forever, as when there are no more rows, do returns "0E0".
I know that this subject has been covered before on PM, see Should string inequality operators return the point the of divergance? and What is truth? (Curiosity corner).
I also recall that the return value of the flip-flop operator .. in scalar context has "E0" appended to signal a terminating match. On some levels, I think that this is very clever, as this is interpreted by perl's string to number parser as "X 10 to the power 0". Hence, using this scalar value in numeric context gives you the line count.
I have several questions in my mind:
Update: I realise with hindsight that I hadn't made it clear that I had solved the original problem. This is the reason it was posted as a meditation, not SOPW.