Create new file per line from txt file, avoid file name too long

Question

I'm creating new file for every line of text in a .txt file

file=/tmp/textFile.txt
while IFS= read -r line
do
        printf "%s\n" "$line" > /tmp/"$line"txt
done < "$file"

I always get an error saying file name too long, is there anyway to shorten the title for the new files to a certain number of characters?

Answer 1

Usig awk to output the contents of the line to a filename with at most 100 characters (preceded by /tmp/, followed by .txt):

awk '{ name = "/tmp/" substr($0,1,100) ".txt"; print >name; close(name) }' /tmp/textFile.txt

If you need to do something about the / and . characters in each name, for example, change them to underscores (this avoids overwriting random files in the filesystem other than the ones in the /tmp directory):

awk '{ name = "/tmp/" substr($0,1,100) ".txt"; gsub("[/.]","_",name); print >name; close(name) }' /tmp/textFile.txt

Answer 2

You can truncate the lines to e.g. 252 characters (which becomes 255 with txt added).

while IFS= read -r line
do
    printf "%s\n" "$line" > /tmp/"${line:0:252}"txt
done< /tmp/textFile.txt

This parameter expansion is from ksh93 and is also supported by Bash and newer versions of zsh, but not portable to sh. The native zsh syntax is $line[1,252] or ${line[1,252]}, the latter also supported by yash.

In any case, as, depending on the filesystem, the limit is often on the number of bytes rather than number of characters¹, the above may not work properly if the text has multibyte characters.

---

^{¹ macos can add further complications in that it may decompose characters such as those with diacritics into a decomposed form storing characters like é as U+0065 and U+0301 instead of just U+00E9 precomposed form.}