"r" == rowe22@commawait6.net <rowe22@commawait6.net>:
r> say i have three files, each contains hundreds of names.
r> now i want to know those names that are in at least two of the files, or
r> find out those in all three files, and put these names in a new file. how
r> do i do that?

Compute the MD5 hash of each line and sort/select using that. Note
that this can get really slow for large files. In such a case use perl
and/or some simpler/faster hash like crc32.

Assuming that your data files are named file1, file2, file3, ..., here
is a simple shell script (let's call it select_common_lines) that does
what you need:

#!/bin/sh
MUST_BE_IN_SO_MANY_FILES="$1";shift
for f in ${1+"$@"}
do
while :; do
IFS= read line || break
echo `echo "$line"|md5sum` "$line"
done < "$f"
done |
sort -k 1.1,32 |
awk '
($1 == prev_line) && (NR != 1){ found_same++;
if (found_same == '$MUST_BE_IN_SO_MANY_FILES') print $3;
}
($1 != prev_line) { found_same=1 }
{prev_line = $1}
'


Sample usage:
$ echo 1 4 5 9 13 17 21 25 29 | tr ' ' '\n' >file1
$ echo 11 13 4 15 17 19 21 | tr ' ' '\n' >file2
$ echo 20 21 4 22 23 24 | tr ' ' '\n' >file3
$ ./select_common_lines 2 file*
4
17
13
21
$ ./select_common_lines 3 file*
4
21