Re: null vs. empty string? (the same old story...)
Posted in 1999
Topics: SQL Development & Query Writing, Triggers, Constraints & Referential Integrity
when selecting from the database you can identify the fields that have been inserted with "" and those that have null by using select from tabname where length(fieldname) < 1 if you don't want to alter your application to insert the keyword NULL instead of "" , then use the above select statement when getting data. At 03:49 PM 5/21/99 +0200, you wrote: >Sorry to bother you all again with the old "null vs. empty string" issue. >I'm sure this has been discussed in this group over and over again but >searching through the archives and other sources I can't find an answer to >my particular problem: > >Informix handles an empty string and the null value differently, i.e. if I >insert an empty string ('' or "") into a string field then that does not >convert to null. This is fine with me as I would expect that behavior from >any RDBMS. However, I now have to port an application from Oracle and SQL >Base, both of which convert an empty string to null when inserting/updating, >to Informix. This application makes intensive use of inserting an empty >string but actually meaning to insert null. > >Is there any way to force Informix to adapt Oracle's behavior of >interpreting an empty string as null? >If not, are there any suggestions for workarounds? I have thought about >writing triggers that would intercept inserts/updates of an empty string and >instead insert/update null, but that seems kind of awkward (besides, I >wouldn't even know how code such a trigger). > >Any help would be greatly appreciated. > >Cheers >E. >