Replace text of unknown content with other text in file

brian0918 · 07-14-2005, 03:19 PM

I need a script that will replace text which I don't know ahead of time on a line that contains words that don't change ("foobar") and are unique to the file. Specifically, I want to make the line that reads:

Code:

          foobar = 123abc                          \

into:

          foobar = xyz456                          \

but I don't know the exact contents of "123abc" ahead of time.

I'll also need to do this with a block of text as well, replacing the block of text between the lines containing "foo" and "bar" with a block of text from a file. Ex:

foo
Text To
Be Replaced
bar

becomes:

foo
Replacement
Text
bar

but, again, I don't know the exact contents of "Text To Be Replaced" ahead of time.

Thanks.

Hko · 07-14-2005, 04:07 PM

Have a look at "sed".
Is this homework?

brian0918 · 07-14-2005, 04:09 PM

Uhh... no. I'm done with the concept of homework. I've been trying to work with sed but no luck.

Matir · 07-14-2005, 04:14 PM

Your second task will require awk, as sed is line-based only.

The first one, you could always just do:

Code:

sed 's/foobar = .*/foobar = xyz456/' FILE > NEW_FILE

There are probably nicer ways to handle this, but that should do it for you, as a quick-and-dirty technique.

brian0918 · 07-14-2005, 04:15 PM

Actually, that doesn't work, because it deletes the \ from the end of the line.

Matir · 07-14-2005, 04:26 PM

change the single quotes to double quotes and insert the variable.

Code:

sed "s/foobar = .*/foobar = ${I_AM_A_SHELL_VARIABLE}/" FILE > NEW_FILE

The braces, though not strictly required, are good for readability there.

They're also especially good if you want this to work

Code:

$ export VAR="test"
$ echo $VARing

$ echo ${VAR}ing
testing

brian0918 · 07-14-2005, 04:27 PM

Actually, that doesn't work, because it deletes the \ from the end of the line.

Matir · 07-14-2005, 04:32 PM

This one should be obvious.

Code:

sed "s/foobar = .*/foobar = ${I_AM_A_SHELL_VARIABLE} \\/" FILE > NEW_FILE

brian0918 · 07-14-2005, 04:43 PM

I got it working by replacing \\ with \\\

Matir · 07-14-2005, 04:47 PM

Oops, sorry. I guess the shell did a bit of interpretation due to the double quotes.

brian0918 · 07-14-2005, 04:58 PM

The most complicated thing I've done with awk is printing a specific column. I have no clue how to replace blocks of text in that way.

brian0918 · 07-14-2005, 05:22 PM

I think I got it to work with sed.

Matir · 07-14-2005, 05:25 PM

Even the bigger block? How'd you do that?

brian0918 · 07-14-2005, 05:29 PM

Found this on linuxjournal:

Quote:

Now, we can combine these last two commands to replace a block of text in a file with the contents of another file. We do it like this:

sed -e '/START/r bar' -e '/START/,/END/d' foo

This finds a line containing START, deletes through to a line containing END, then reads in the contents of the file bar. Because the r command doesn't read in the file until the next input line is read, the d command is executed before the new text is read in, so the d command doesn't delete the new text, as one might expect, looking at this command. The -e option tells sed that the next argument is a command, rather than an input file. Although it is optional when there is only one command, if we have multiple commands, they must each be preceded with -e.

http://www.linuxjournal.com/article/1224

vladmihaisima · 07-14-2005, 05:35 PM

And here is an awk solution:

Code:

 awk -- " /foo$/ {p=1;}; /bar$/ {e=1;}; /.*/ { if(p==1) {print \"somethingelse\nsecond row\"; b=1; p=0; }; if(b!=1) { print; }; if(e==1) { e=0; b=0; }}; "

Not as clear as sed though, but that comment about

Quote:

so the d command doesn't delete the new text, as one might expect, looking at this command

tells me maybe I don't like very much sed for a good reason. I admit that the above command is not very clear, but for someone used with flex, it's not that bad.(and it does what one might expect

)