SED, regexp or such

aolong · 03-07-2008, 02:23 PM

I have a long list of names in mysql field 'first':

Andrew H.
Mary
John T.
etc...

And a lot of fields in 'last' like:

Huges, MD
Rowe, DDS, CPR
Bonner
etc...

I am trying to clean this up to remove everything after the first space in 'first' and everything after and including the comma in 'last'.

I have failed with mysql regexp and thought to export the names and try sed. Can anyone help with this?

Andrew

Wells · 03-07-2008, 02:29 PM

Well, if you have them exported it is probably easier to use awk on them:

Code:

$ cat firstnames_list.txt | awk '{print $1}' > new_firstnames_list.txt
$ cat lastnames_list.txt | awk -F , '{print $1}' > new_lastnames_list.txt

The "-F ," portion in the second one tells awk to split on commas instead of on spaces, which is the default. You might have to run it through a couple of these filters to get rid of all of it.

If you wanted to print the second (or third, or whatever) item that is split by the token you have chosen (spaces, commas, whatever) just change the $1 to $2 or $3 or whatever number you want.

Tinkster · 03-07-2008, 02:30 PM

What does the exported file look like? Can't make a guess
as to a valid regex if I don't know the input.

Cheers,
Tink

jschiwal · 03-07-2008, 02:32 PM

Are these two lists or are you showing to types of patterns from one list?
sed 's/ .*//' filename
sed 's/, .*//' filename2

Or if from the same list you can have two sed commands on one line:
sed 's/ .*//;s/, .*//' filename
Another form is the -e option:
sed -e 's/ .*//' -e 's/, .*//' filename

aolong · 03-07-2008, 02:36 PM

Quote:

Originally Posted by Wells

Well, if you have them exported it is probably easier to use awk on them:

Code:

$ cat firstnames_list.txt | awk '{print $1}' > new_firstnames_list.txt
$ cat lastnames_list.txt | awk -F , '{print $1}' > new_lastnames_list.txt

The "-F ," portion in the second one tells awk to split on commas instead of on spaces, which is the default. You might have to run it through a couple of these filters to get rid of all of it.

If you wanted to print the second (or third, or whatever) item that is split by the token you have chosen (spaces, commas, whatever) just change the $1 to $2 or $3 or whatever number you want.

Man, that was fast and it works like a charm!

Thanks!

pixellany · 03-07-2008, 02:36 PM

If you indeed have the data the way you show it, it is easy:

For the "first" list---remove all after the " ":

sed -i 's/ .*//' filename (deletes everything from the first space to the end of the line the regex reads literally as "a space, followed by any number of characters")

A similar approach will work for the second case....