Transpose the entries

iqbala has asked for the wisdom of the Perl Monks concerning the following question:

Hi Monks

I am trying to get the following input in transpose form

---------------
Name: Tim Brown
username: txbro27
cuid: twbrown
---------------
Name: Dean Walker
username: dxwalk3
---------------
username: jxrizzo
Name: John Rizzo
---------------
Name: Doreen Mirlass
username: mirlasd
---------------
cuid: chamblem
Name: Mike Chambless
username: mblem
---------------
name: Tim McCarthy
username: txmccar
cuid: tmmccar
---------------
[download]

There are three things to notice here

- The order of the Name/cuid/username not always same.
- They need to considered as case insensitive
- sometime the cuid is missing
The output should be in this format

username:fname lname:cuid
[download]

If the cuid is missing the value should be EMPTY. For
example the above input will give the following output

txbro27:Tim Brown:twbrown
dxwalk3:Dean Walker:EMPTY
jxrizzo:John Rizzo:EMPTY
mirlasd:Doreen Mirlass:EMPTY
mblem:Mike Chambless:chamblem
txmccar:Tim McCarthy:tmmccar
[download]

Any help with this would really help me a lot
Thanks

Comment on Transpose the entries Select or Download Code

Replies are listed 'Best First'.
Re: Transpose the entries by boo_radley (Parson) on Jan 08, 2004 at 15:14 UTC
First off, I'd set the record separator to that big line of dashes, then push each 'line' into an array. I'd do this in a block, so that the record separator would automatically return to it's regular value of "\n" after the block. Then I'd go through the array I created in step one, and create a hash by splitting on ": " and newlines, taking care to make sure that a 'cuid' key exists. I'd push this hash into a different array. At that point, I'd be ready to loop through this second array to create output in whatever format I'd need. update What abigail-II says.	[reply]
Re: Transpose the entries by borisz (Canon) on Jan 08, 2004 at 15:21 UTC
Like this? #!/usr/bin/perl local $/ = "---------------"; while( <DATA> ) { next if /^\S$/; my ( $n, $u, $c ); ( $n ) = /Name:\s+(.)$/mi; ( $u ) = /username:\s+(.)$/mi; ( $c ) = /cuid:\s+(.)$/mi; $n \|\|= 'EMPTY'; $u \|\|= 'EMPTY'; $c \|\|= 'EMPTY'; print "$u:$n:$c\n"; } __DATA__ --------------- Name: Tim Brown username: txbro27 cuid: twbrown --------------- Name: Dean Walker username: dxwalk3 --------------- username: jxrizzo Name: John Rizzo --------------- Name: Doreen Mirlass username: mirlasd --------------- cuid: chamblem Name: Mike Chambless username: mblem --------------- name: Tim McCarthy username: txmccar cuid: tmmccar --------------- [download]	[reply] [d/l]
Re: Transpose the entries by Abigail-II (Bishop) on Jan 08, 2004 at 15:14 UTC
Here's a general outline of the solution: Open the file. Read in a record. Process the record. Print the processed record. Finish if you have read in the last record. Go to step 2. Abigail	[reply]
Re: Transpose the entries by Fletch (Bishop) on Jan 08, 2004 at 15:11 UTC
Fish. `my %rec = ( cuid => "EMPTY" ); scalar <>; # eat first seperator line. while( <> ) { chomp; if( /^-+$/ ) { print "$rec{username}:$rec{name}:$rec{cuid}\n"; %rec = ( cuid => "EMPTY" ); } else { my( $k, $v ) = split( /:\s/, $_, 2 ); $rec{ lc $k } = $v; } }` [download] Update*: Missed the part about the default.	[reply] [d/l]
Re: Transpose the entries by Theo (Priest) on Jan 08, 2004 at 15:11 UTC
What have you tried so far? -Theo- (so many nodes and so little time ... )	[reply]