Re: Writing to csv file

"CSV" or Comma Separate Value has evolved. If you actually use commas, then there is the problem of commas within quotes. The program and regex logic to deal with with this is actually quite complex.

An alternate to the comma that is used in many data base outputs is the "pipe" or | symbol. Although common in programming languages, this symbol is almost unknown in name, address, department name fields, etc. A database output that uses the "|" instead of the "," to separate fields is also known as a "CSV" file. Microsoft Excel can easily import CSV files separated by "|".

If I have a choice when generating the file, I use "|". When I have to read a file from some other source, very often but not always it uses "," and I have to deal with that.

So, if you generate the file, I would use "|" instead of "," because it makes the parsing easier - which is reason enough.

Comment on Re: Writing to csv file

Replies are listed 'Best First'.
Re^2: Writing to csv file by GrandFather (Saint) on Mar 03, 2011 at 00:27 UTC
Using a different character, or even a string of characters, makes no difference to the nature of the problem of the possibility that the separator may validly be part of the data. The much better solution is to quote the separator string in some fashion if it appears in the data. Such quoting is what CSV modules do, albeit with some lack of standardisation in how the quoting is achieved. True laziness is hard work	[reply]
Re^3: Writing to csv file by maestromani (Initiate) on Mar 03, 2011 at 10:25 UTC
Thanks a lot for all the reply. I'm generating csv for crystal report. If i replace , with \| Crystal report is not identifying a separate column.. All i did is used "=~ s{,}{::comma::}g;" as advised by John BRs	[reply]
Re^3: Writing to csv file by Marshall (Canon) on Mar 05, 2011 at 02:09 UTC
I don't see any real disagreement here. Some of the databases I work with use "\|" instead of "," and disallow "\|" as a valid data character - so this "\|" within "\|" issue does not come up. When I generate a CSV file on my own, sometimes I use the "\|" instead of "," when I know that "\|" is not going to be in the data. I'm just saying that option is possible. Although the format is called "Comma Separated Value", the character can be anything you want, which is what you are saying. Some folks take CSV literally and don't realize that some other character can be used and that standard programs can parse files like that. To parse the general case of a CSV file, I would use one of the Perl modules as this is a very deceptively easy looking, but very hard to implement specification.	[reply]