Reputation: 2101

comparing files

hi guys i have two pipe delimited files,first file contains 1000 records and second file contains 10 records and records which are present in second file exists first file. Now need a file which which will give me 990 records by excluding 10 records which occur in both files,

i know how to this using SQL .But how can we acheive this in UNIX?

Any help greatly appreciated

Upvotes: 1

Answers (5)

ghostdog74

Reputation: 342333

awk 'FNR==NR{ a[$0++;next}(!($0 in a))' some all

Upvotes: 0

John Kugelman

Reputation: 361585

First, replace the pipes with newlines so we can process the files with line-oriented Unix tools. Then use comm to filter out the lines that appear in both files. Finally, convert the newlines back to pipes.

tr '|' '\n' < file1 | sort > file1.sorted
tr '|' '\n' < file2 | sort > file2.sorted
comm -3 file1.sorted file2.sorted | tr '\n' '|'

Or done all on one line using process substitution <(command) syntax:

comm -3 <(tr '|' '\n' < file1 | sort) <(tr '|' '\n' < file2 | sort) | tr '\n' '|'

Upvotes: 0

Ken Keenan

Reputation: 10538

The comm utility can be used to filter lines from one file out of another

Upvotes: 0

Martin v. Löwis

Reputation: 127447

Suppose the files are all and some. Then do

 fgrep -v -f some all

Upvotes: 8

Pierre

Reputation: 35236

see the join command

Upvotes: 1

comparing files

Answers (5)

Related Questions