Shell Programming and Scripting

Hi,

I need a shell/command to achieve this task.

I've a delimited unloaded file from oracle in a scrambled format as shown below with many blank lines in it, I'm just trying to tailor it in a format that would be compatible to view and load it to a IDS db.

Here is the problem

Infile-

asd,sdfgsdf,,
<blank line 1>
fgsdfg, , , ,asfa,
<blank line 2>
<blank line 3>
sdfd,sdfd, ,
asd,asdd,
asdf, , , ,dsfd,
afdd,dsd

blank line 1 - potential data
blank line 2 - potential data
blank line 3 - record seperator

In the above ex, blank line1,2 is not a record seperator but it has a NULL data which is identified in the ASCII format. I'm trying to make the file into each data record in a single row as below, replacing the non-viewable data fields into some string (*$& here)

Output-
asd,sdfgsdf,, *$&,*$& ,fgsdfg, , , ,asfa, *$&
sdfd,sdfd, ,asd,asdd,asdf, , , ,dsfd,afdd,dsd

Please note that each blank line can have one or more fields in it

Thanks

I'd try to look at those blank lines a little closer. Try to pipe them to 'od' : they should differ, if they contain different whitespace characters.
Then, you could use e.g. sed to replace the white chars with something else:

Code:

[MIRO@stroke]$ echo -e "\r"  | od -x
0000000 0a0d
0000002

[MIRO@stroke]$ echo -e "\r" | sed 's/\x0D/blah/g'
blah