Re^2: Data managing problem

[tmharish]

>perl -anle"$h{ $F[1] } = $_; }{ print for values %h"

I understood that about the implicit split and the -loctnum after reading this ( which is awesome by the way )

>perl -anle"$h{ $F[1] } = sprintf qq[%05u%s], $., $_; }{ print substr $_, 5 for sort values %h"

I have been searching around for a bit and I simply dont get that - could you explain that one please?

[BrowserUk]

Sure

First the switches: perl -anle"..."

• -e"...": a small program to run immediately
• -n: Feed the data (from the file(s) listed on the command line; or STDIN), to teh -e snippet above on line at a time in $_

  This has the effect of placing the -e"..." within a while loop very similar to this:

  while( <> ) {
      ... # the -e snippet here
  }
  

• -l: auto chomp those lines on input; add a newline to print statements on output.
• -a: autosplit those lines on whitespace (or the argument to -F if used) and place the results in @F

Now the -e snippet code in three parts:

1. "$h{ $F[1] } = sprintf qq[%05u%s], $., $_;

   This builds %h with the second field of each input line as the key; and the whole line ($_) prefixed with the line number ($.) as the value.

   Later lines with matching second fields will overwrite earlier ones.

2. }{

   This has the effect of converting the while loop shown above into:

   while( <> ) {
       ... # the part of the -e snippet before the }{ here; runs for ever
   +y line in the file.
   }{  # closes the while loop body and puts a bare block after it
      ... # The part of the -e snippet after the }{ goes here and is only
   + executed after the input is exhausted.
   }
   

   It a little like adding an END{} block containing the second half of the -e snippet.

3.  print substr $_, 5 for sort values %h"

   One we've read all the input, only the last line with each second field value will remain in the hash, prefixed with the line number it came from.

   So sort the values from %h to get them into line number order and then print them out having removed the line numbers.

[tmharish]

Thank you very much for the detailed explanation.

Couple of follow up questions if you dont mind - Just to make sure I am understanding this right:

1. The printf is used to ensure that should the line numbers jump from single digit to two ... this has the ability to deal with up to 5 so we can substr accordingly - is that accurate?
2. Does the sort at the end assume that the input ( and the required output after removing rows ) is sorted or am I missing some way in which the ordering of the original input data is maintained?

[BrowserUk]

The printf is used to ensure that should the line numbers jump from single digit to two

The sprintf pads the lines numbers with leading zeros so the sort correctly. Ie so that they do not sort as:

1
10
11
...
2
20
21
22
...
3
30
31
32
33

%05u means we can correctly sort files with upto 99999 lines. If you need more change the number in the format.

Does the sort at the end assume that the input ( and the required output after removing rows ) is sorted

Why would sort "assume its input was sorted"? We are sorting them because we know they will not be. That's why we added the line numbers so that we can put the, back into the input ordering. Perhaps the following where I've left the line numbers in place will clarify things?

C:\test>perl -anle"$h{ $F[1] } = sprintf qq[%05u%s], $., $_; }{ print 
+for sort values %h"
1/2/2013     cgoo     nreuiheru
1/4/2013     doow    reiqrqueih
1/5/2013     hellio    ruieqrhfuepqh
1/20/2013    cgoo    3rhquh4ureyh
1/30/2013    yetil     jerqohgqrij
2/13/2013    hellio   rueqipheruh
2/14/2013    cgoo    wehrig4r74378
^Z
000021/4/2013     doow    reiqrqueih
000051/30/2013    yetil     jerqohgqrij
000062/13/2013    hellio   rueqipheruh
000072/14/2013    cgoo    wehrig4r74378

[Anonymous Monk]

see perlrun for most recent version

[tmharish]

Thank you!