Carriage return characters within string variable in txt format

[Adrien Bouguen]

Dear statalister,

I have a weird problem while insheeting data from a txt file. For some reason, one of the string variable contains a CR (carriage return) characters that makes the line breaks and mess up my import. At the end of each observation, there is a CR and LF character.

first line: Exclusion de cours;;15/09/2014;Dissipation, cr
Manque de respect caractérisé; cr lf
second line: Recopier le cours p. 16 et 17, le savoir faire p. 18 et 19: 1, 2, 3. cr lf

If I simply insheet this txt file the first observation is diplayed in two lines.

One way I found to deal with the problem is to remove all cr from the txt file using filefilter:
filefilter "EXP.txt" "EXP_clean.txt", from(\r) to("")

It works fine but since I am going to use this problem in thousands of txt file I am affraid it is going to mess up others by removing all cr.

Have you ever experienced something like that?

Thanks so much

Adrien

[Sergiy Radyakin]

Adrien,
you can use an intermediate substitute to avoid the problem.

do http://radyakin.org/statalist/2015/20150319_1096702.do

Best, Sergiy Radyakin

[Mike Lacy]

I'd also note, per an earlier post of mine , that -fileread-() and -filewrite()- can be nicer than -filefilter- for a purpose like this. The approach is the same as Sergiy's, but -subinstr()- is used to clean up a strL that contains the whole file that
needs cleaning.

Regards, Mike