UNIX for Dummies Questions & Answers

Hi,

To comply with a new naming convention on a mediawiki site we have to run a SED or other PERL command to change all instances of

[[Xxxxx xxx Xxxxx]] or [[Xxxxx Xxxx]] or [[Xxxxx Xxxxxx Xxxxx]]

to

[[Xxxxx xxxxx xxxx]]

Can someone please explain how to do this...

It has to be done on a mysql dump, so if there is a way to do this in mysql even better, otherwise I will make a dump of the database and run it then and restore the database.

Thank you.

---------- Post updated 22-11-09 at 05:05 AM ---------- Previous update was 21-11-09 at 06:59 PM ----------

Pleaseeeeeeeeeeeeeeee

Assuming [[Xxxxx xxx Xxxxx]] or [[Xxxxx Xxxx]] or [[Xxxxx Xxxxxx Xxxxx]]
means letters A-Za-z I do not see how this can be done without a lot more information

For example:
[[Xxxxx xxx Xxxxx]] becomes [[Xxxxx xxxxx xxxx]]. The middle term gained characters xxx to become xxxxx. Where do the extra characters come from? The last term lost one character Xxxxx to become xxxx. Do we drop the X?

your OR examples aren't the same length. The new string is len 14, x's only, and the original 13, 9 and 16 . So you need to define rules of how to handle over and under size as most databases don't necessarily handle overs well.

in sql you could do it but it would be messy but look up the equivalent of charindex, Sybase, in mysql and strip the spaces then substring the result out into the new format applying whatever rules you decide for overs and unders.

I'm assuming the [ is for display purposes? If not then it's not an issue just ignore them.

dump to file, assuming a one field file.
awk -F" " '{print tolower($1$2$3)}' tmp.tmp > tmp.tmp1 #strips spaces
awk '{print toupper(substr($0,1,1)) substr($0,2,4) " " substr($0,5,5) " " substr($0,11,5)}' tmp.tmp1 > newfile.tmp

if you read up on BEGIN and END you can string it all in one awk statement. However theres more work needed to handle the unders and overs.

You could do it like this awk -F " " '{print tolower($1$2$3)}' tmp.tmp | awk '{print toupper(substr($0,1,1)) substr($0,2,4) " " substr($0,5,5) " " substr($0,11,5)}' > newfile.txt but I usually prefer to do each command on a separate line so when I'm trying to work out what's not working at 2am I don't have to parse the piping of commands.