Shell Programming and Scripting

Hello,
i have a file like that

Code:

foo A new
bar A new
bar B new

I need to replace 'new' with 'done',
but only in lines containing 'bar' AND 'A'.

output file should then become

Code:

foo A new
bar A done
bar B new

Sorry im not able to figure it out, not even shure if i should take sed.
Thanks a lot for suggestions.
knoxo

Sounds like an awk problem actually. But sure, it can be done in sed, too.

Code:

sed 's/bar A new/bar A done/' filename

Thank you for your reply, but i forgot to mention that i can not rely on the order of 'A' and 'bar'. i just can assume that they appear anywhere in that line.
So perhaps a modification of this sed-command would do the task.

For example,

Code:

sed '/bar A/ s/new/done/' filename

would search for the pattern "bar A" and then replace new with done in that particular line.
But what I need is an expression for the pattern like
"'bar' AND 'A' appear in that line".
Now its more like
"'bar A' appears in that line".
Anyone who can help?

Code:

sed -e '/bar/!n' -e '/A/!n' -e 's/new/done/' file

This skips to the next line if there is no match on bar, and then if there is no match on A.

You might want to tighten up the expressions since a stray A in the middle of another word would match the second condition, but without more information, we can't help you with that. If you can be sure there will be a space on both sides then that's obviously a trivial improvement which would reduce the risk of false matches somewhat significantly.

Great, that works!
'A' and 'B' were just examples, they actually correspond to longer strings...

Perhaps i should change the syntax to prevent such kind of matches if the search text appears anywhere in the line.
What if the file consists of colums instead (column delimiter is ':'):

Code:

foo : A : new
bar : A : new
bar : B : new

how could i restrict the search and replace condition to

"if column1 contains 'bar' AND column2 contains 'A', then replace any text in column3 with 'done'.
knoxo

Try:

Code:

awk 'BEGIN { FS=OFS=":" } $1 == "bar" && $2 == "A" { $3="done" }' inputfile

Jean-Pierre.

hello Jean-Pierre,

edit:

Now it works, i had additional blanks around the 'A' and 'bar', which is different from ' A ' etc so no math was found.
How could i make it insensitive to leading or trailing blanks?

However,

Code:

awk 'BEGIN { FS=OFS=":" } 1==1 { $3="done" }' inputfile

does not give any output either, while

Code:

awk 'BEGIN { FS=OFS=":" } 1==1' inputfile

prints all the contents of the file.
Help me figure out whats happening here?
knoxo