[prev] [thread] [next] [lurker] [Date index for 2006/12/18]
* Earle Martin <hates-software@xxxxxxxx.xxx> [2006-12-18T09:06:34] > "Foo","Members List Report","00003816","Somebody Incorporated23B > Snibbits Building69 Foonly StreetLondonFO0 8AR+44 20 7123 4567","a > couple","of other", "fields here","Report Run Time: 05 Dec 2006 at > 10:53:50","Page -1 of 1","Company Name Snipped ","Enterprise MRM" At a former job, we had a customer who demanded CSV manifests of all our shipments. They had multiple business units, each of whom used a slightly different spec for the output. The output spec would change from time to time, too, and there was no revision number. If they re-ordered an old part, the order would include the old format spec. It was a crapshoot whether they actually wanted the old spec. Changes included things like, "change date from YYMMDD to MMDDYY." I'm sure that whatever they loaded those files into was REALLY reliable data. The real hate for me, though, with CSV isn't asinine customers, but CSV itself. People say "a CSV" file as if that were a reference to some well-understood format. No two things seem to agree on whether a set of values can include commas, newlines, escaped quotes, or what. Oh, and how do you escape quotes? \" or "" or something even stupider? Also, check out this awesome comment found via Google: As is the case with most exchange formats since XML, CSV files have become somewhat of a legacy format. -- rjbs
Generated at 00:01 on 20 Dec 2006 by mariachi 0.52