UNIX for Dummies Questions & Answers

Hi All,
I am having the below sample lin in the file.

Code:

'''||V_D_BOOKING_TYPE||''' || ''¬'' || '''||V_D_SHARE_DELIV_FL||''' || ''¬'' || '''|| SUBSTR(V_D_FINANCIALDATATYPE,1,1) ||''' || ''¬'' ||

I need to replace ''¬'' with char(172). how do i do it.
PLease help.

Code:

sed 's/\''¬\''/chr(172)/g' test.txt  # not correct.

Try:

Code:

sed "s/'¬'/chr(172)/g" test.txt

I'm a bit surprised: ASCII chr(172) (decimal 172) IS¬. What your purpose of replacing it? Or do you mean the literal string "chr(172)"?

The sed command I suggested assumes that abhi_123 wanted the literal string "chr(172)". The byte with decimal value 172 is not as ASCII character at all. (ASCII only defines bytes with values 0 through 127 and a character with the glyph ¬ is not one of them.) The EBCDIC character representing the glyph ¬ has decimal value 95. The ISO/IEC 8859-1 glyph for the character with decimal value 172 is ¬. The 8859-2 glyph for the character with decimal value 172 is Ť. The 8859-3 glyph for the character with decimal value 172 is Ĵ...

Probably more important, however, is that in UTF-8 (which is becoming more common as a code set AND appears to be the encoding used in the input abhi_123 supplied) the two byte character representing the glyph ¬ has bytes with decimal values 194 and 172 (in that order) and the single byte with decimal value 172 is an encoding error.

If abhi_123 is trying to convert an input file encoded in UTF-8 into a file encoded in 8859-1, the iconv utility should be used instead of trying to map an individual character's value into an invalid character in the output data stream being written by sed.