convert CSV into another format

[alexpank]

Yup, that's one seriously strange guy... Still, I guess you'd have to be to call yourself 'why the lucky stiff'. Oh well, to each their own, diff'rent strokes for diff'rent folks, it takes all kinds to make a world and all that.

 

BTW (kinda OT) I've got a whole stack of CSV's that I'm looking to convert into another format. Should I use python or ruby or something like that, or is there a better alternative?

[theYinYeti]

Splitted from https://mandrivausers.org/index.php?showtopic=26192 so as this topic not to go off-topic :) Yves, moderator.

[neddie]

perl? sed? OpenOffice Calc?

Depends a bit on what your "other format" might be, but if you can describe your changes with regular expressions then sed is probably one of the easier ways to go.

[iphitus]

Just depends on what you are converting to.

 

If you need to use a programming language of some sort, I guess python wouldnt be too bad, it's even got a csv read/write module built in. http://docs.python.org/lib/module-csv.html

[Hirogen2]

Text::CSV http://search.cpan.org/~alancitt/Text-CSV-0.01/CSV.pm

You could transform a CSV into a '\0' separated using something like

perl -MText::CSV -lne 'INIT{$c=new Text::CSV}$c->parse($_);print join("\x00",@{$c->fields()})}'

..Untested..

[alexpank]

Ta for splitting that off, Yves. I know, I know, I should've done it myself *flinches in preparation for beating*

 

To give a bit more detail, I've got a bunch of CSVs from MS (yes, I have sold my soul to the devil), which are basically the text strings from most of their programs, with the English and Japanese versions in separate fields. What I want to do is convert those CSVs into something like

 

 [I]Japanese text[/I] /English text/ (file that it came from)

 

-- the formatting's to do with a dictionary program I use. There are a number of other fields in the CSV, but i was just going to throw them out.

 

So, given that info, is python still a good option?

 

TIA (and apologies for the OT post...)

 

Alex

 

PS (more OT) MS puts out these files so that they can 'ensure consistency in translation' for third-party programs, but I've found countless situations where, for no apparent reason, their own files are inconsistent! Not that this suprises me - MS making dodgy software? :o

[neddie]

If the input file is of the form

english,japanese,other,ignored,fields

then you could convert all the commas to spaces with something like this:

sed "s/,/ /g" inputfile.csv

To convert to the kind of format you want, you could do something like this:

sed "s/^\([^,]*\),\([^,]*\),.*$/\[i]\2\[\/i\] \/\1\//" inputfile.csv

where in this case the first two fields are swapped with a [ i ] and [ /i ] around the japanese and / and / around the english. I'm not quite sure how to get the filename but that should get you started.

[alexpank]

Thanks for that, neddie. I've been reading a little bit about gawk, and I might see if that can do it too. I have a feeling that the filename (or something liek that) might be one of the fields, so I could get around it that way.

 

Thanks for your help!

 

Alex