Shell Programming and Scripting

I have files starting with some char 'X' (they could vary from day to day) and I have to read all the files starting with char X*.txt and write to a single file Y.txt. But I have to omit the first line in each file (header) and consider only the header in first file i read.

After all the writing is done to Y.txt I should sftp it to a server.

Appreciate your help regarding this.

Thanks.

hello ,

Please rephrase this - "But I have to omit the first line in each file (header) and consider only the header in first file i read."

If the number of files is small enough to glob properly, you can do

Code:

tail -q -n +2 X*.txt > Y.txt

To get the header of the first file:

Code:

for FILE in X*.txt
do
        read LINE < $FILE
        echo "$LINE" > Y.txt
        break
done

tail -q -n +2 X*.txt >> Y.txt

..presuming the number of files doesn't change between the first execution and the second.

Lets say the files are like this

File Name: X1.txt

header1st
abc
bcd
age

File Name: X2.txt

header1st
abcege
ageerw
ageaer

.......

in my result file Y.txt

I should have only

header1st
abc
bcd
age
abcege
ageerw
ageaer

I don't need to add all the headers of every file I read, I need only the first file (which ever I read first) header, since all the headers are going to be the same.

Here's a way that won't mess up when the number of files changes.

Code:

# Open FD 5 for writing to Y.txt
exec 5>Y.txt

TMP=$(mktemp)
ls X*.txt > ${TMP}
# Open FD 6 for reading from the temp file
exec 6<${TMP}

read FIRST <&6
cat "${FIRST}" >&5

xargs tail -q -n +2 <&6 >&5

# close streams
exec 5>&-
exec 6>&-

This is untested.