I have text files of several thousand rows and I need to append a string (.php) at the end of the first line of all of them

I've managed to understand the sed program to the point of writing a working command for stdout:

sed -e '1,1s/$/.php/1'

It works great if I write the output files to a different directory (made for the occasion):

#!/bin/bash
for X in *.txt
do
sed -e '1/1s/$/.php/1' $X >> ~/appended/$X
done

If I write $X >> $X, the filenames of the current directory remain but their text is all gone.

How can I output the file with the same name to the same directory?

i run in to this problem a lot to. im not sure what the clean way to do it is, but an ugly fix is to:
#! /bin/bash
for X in *.txt
do
sed -e '1/1s/$/.php/1' $X > tmp
mv tmp $X
done


you only want to use one '>' so it will overwrite tmp everytime instead of appending to it.

If your sed is version 4.x use the -i flag.

Tinkster,
#!/bin/bash
for X in *.txt
do
sed -e '1/1s/$/.php/1i' $X > $X
done

emptied the output *.txt files. Does w for write prevent this?

sirclif, the tmp trick works as expected . When i tested it, it took some time to move all the huge files. I guess that mv parses all the file length or the memory reads through all over again. I wonder if there is a way of making the bash edit the first line and go.

sed -i -e ....
[edit]
-i stands for in-place, no redirection at all required.

Read the man page, quite good reading.
[/edit]


Cheers,
Tink

#!/bin/bash
for X in *.txt
do
sed -i -e '1/1s/$/.php/1' $X
done

did the job. Thanks!

Also, advanced reads:

Manipulating Strings with sed http://www.unet.univie.ac.at/aix/aix...trings_sed.htm

sed tutorial
http://pegasus.rutgers.edu/~elflord/unix/sed.html

handy one-liners for sed
http://www.student.northpark.edu/pem...d/sed1line.txt

executable sed script
end of http://www.cs.hmc.edu/tech_docs/qref/sed.html

oh wow, that is a handy thing to know.

Tinkster, your sed sorcerer. I always see you giving good tips on sed.