UNIX for Dummies Questions & Answers

Hey guys.

I am trying to sort a file by account number through UNIX. I have a few things but it seems to sort by account number AND sort everything after the account number.

Help please.

Thanks

If you show what the input looks like and what the desired output should be then it would be easier to suggest something.

You said: "AND sort everything after the account number"
What do you want to have happen when similar records are found?

input=
002x
003d
001a
003a
001b

output from sort on acct # (first 3 positions)
001a
001b
002x
003a
003d

The underlying problem seems to be that whenever it finds a match when its sorted. It compares additional data after the account number to see if they are equal. I want the sort to just ignore these two same account numbers and continue normally without looking at data past the account number field.
So in other words, how can I just ignore the two account numbers that match without having it look past the acct nbr?

Hope that makes sense.

What do you mean by that? A sort re-orders data, so are you asking to 'delete' records that are not unique?

This is becoming confusing... so please just paste in some sample data.

Run:

Code:

man sort

Check out the "-k (--key)" option. You can have it sort by just one field.

Here's an example of sorting by the second column of a comma-separated file:

Code:

 $ sort -t "," file.dat -k 2

If your file is currently separated by whitespace, you can ignore the whole -t bit and it should work just with the -k option.

Shawn