can shell script do this job pls？

BJTD

各位大侠，bjgirl MM大侠：

bash, rh7.3

problem:

1). file f1 and file f2
2). f1 1000 long lines
3). f2 50 short lines
4). f1 contains f2
5). remove all lines in f1 which contain the short lines in f2

i know i can write a small c application to do this combining with sed.

can shell script do this job pls ？


thx in advance

BJTD

one problem with this solution (cat f1 f2 |sort| uniq -u) is that:

if f1 has dulicated lines which contains a line in f2.

(maybe do the command a few times)

BJTD

原帖由 "Mercury_cn" 发表：

这句话好费解,C还是SED?

forgive my poor english

i meant calling sed/awk from my c application~~

BJTD

原帖由 "luvlm" 发表：
uniq seams can not do the job.

reason: lines in f2 are not necessarily the same as in f1

if you gurus read the original poster carefully.

indeed.  however, you can specify parameter -w,
which can reasonably do the job in my situation.


thank you all!

luvlm

uniq seams can not do the job.

reason: lines in f2 are not necessarily the same as in f1

if you gurus read the original poster carefully.

Mercury_cn

[quote]原帖由 "BJTD"]i know i can write a small c application to do this combining with sed. ....[/quote 发表：


这句话好费解,C还是SED?

ailms

用uniq和sort就可以实现了

如花

AIX中有个很有趣的命令，不知道其他系统有没有
comm [ -1 -2 -3 ] File1 File2
-1 Suppresses the display of the first column (lines in File1).

-2 Suppresses the display of the second column (lines in File2).

-3 Suppresses the display of the third column (lines common to File1 and File2).
使用举例：
[code]
cat aa
111
222
333
cat bb
111
222
comm -3 aa bb
333
[\code]

BJTD

原帖由 "yichun" 发表：
如果不考虑顺序的话,可以:
cat f1 f2 | sort | uniq -u

thank u very much indeed.

i shall give it a go later on.

have a good weekend, all.