[SOLVED] awk for cutting off the snake's head :)

tabbygirl1990 · 04-16-2014, 10:16 AM

hi guys,

i need a little awk data manipulation help...

i have two files as input files (snake.dat and knife.dat)and i need an awk script to output the head.dat file

Code:

snake.dat
1   1.71
2   8.02
3   1.03
4   2.04
5   3.15
6   4.06
7   5.07

Code:

knife.dat
5   1.54
6   8.06
7   1.23

and the desired output file would be

Code:

head.dat
5   3.15
6   4.06
7   5.07

so my "thinking algorithm" is to start with the knife.dat file, read the first row first column value (in this example 5). then use that value of 5 as a "look-up key" in the first column of the snake.dat file, and then use the look-up key (in this example 5) as a knife to cut the head off the snake.dat file and output just the head of the snake, the file head.dat

thanks for whatever help!

tabby

TB0ne · 04-16-2014, 10:51 AM

Quote:

Originally Posted by tabbygirl1990

hi guys,
i need a little awk data manipulation help... i have two files as input files (snake.dat and knife.dat)and i need an awk script to output the head.dat file

Code:

snake.dat
1   1.71
2   8.02
3   1.03
4   2.04
5   3.15
6   4.06
7   5.07

Code:

knife.dat
5   1.54
6   8.06
7   1.23

and the desired output file would be

Code:

head.dat
5   3.15
6   4.06
7   5.07

so my "thinking algorithm" is to start with the knife.dat file, read the first row first column value (in this example 5). then use that value of 5 as a "look-up key" in the first column of the snake.dat file, and then use the look-up key (in this example 5) as a knife to cut the head off the snake.dat file and output just the head of the snake, the file head.dat

That's certainly one way to do it. Since you need help with your script, why don't you post what you've written/tried so far, and tell us where you're stuck. We're not going to write your scripts for you, though, but will be happy to assist.

grail · 04-16-2014, 11:28 AM

Agreed ... your solution "sounds" like it has merit but you will need to post solution and where you are stuck if you want us to help

allend · 04-16-2014, 11:31 AM

A nasty little pipeline, rather than elegant awk.

Code:

cat knife.dat | colrm 2 | sed 's/^/\^/' | grep -f - snake.dat

tabbygirl1990 · 04-16-2014, 02:07 PM

oh sorry guys, i had to run off i didn't mean for you all to write it

here's what i've been stabbing at...

BUT!!! in what i've written below, i did one major thing wrong. the knifevalue var should be "looked-up" as the first value in the first column of knife.dat, not user specified (set to 5) as i did below (please don't laugh

i don't know how to do that, look up and pass over in a var field.

Code:

awk -knifevalue=5 -F " " 'FNR==NR{if($1==knifevalue){knifevalue==$1;nextfile};next}$1==knifevalue{print $0}' snake.dat knife.dat > head.dat

ok so here's my thinking and 1st attempt to write something

tell awk to look in the 1st column of the first file (snake.dat) for the knifevalue that i told it, on the same line as knifevalue is found in the "nextfile" the second file (knife.dat), when it occurs (it's equal) then print that entire line (there's actually more columns of data than I showed in the example files) to the head.dat file

metaschima · 04-16-2014, 02:27 PM

Try:

Code:

bash-4.2$ awk 'FNR==NR{ if (NR == 1) knife=$1; next }{ if ($1 >= knife) print }' knife snake
5   3.15
6   4.06
7   5.07

grail · 04-17-2014, 12:01 AM

If i may, is it just a coincidence of data presented that the numbers in the first column are from both files? ie do you want all that match from knife and found in snake to be shown
or only the first number that matches? Also, what happens, assuming current example, if '5' is not in snake but '6' is?

Neither of the current solutions allow for this second scenario.

tabbygirl1990 · 04-17-2014, 10:12 AM

good morning grail, and the other guys.

yea i just made up the example data, the real data is on a closed lan, i'd love to post it but i can't

last night i played with the example data some more at home (not when i do my best thinking) and ran into the same kinds of problems.

so a different set of example knife files...

Code:

snake.dat
1   1.71
2   8.02
3   1.03
4   2.04
5   3.15
6   4.06
7   5.07

and

Code:

knife.dat
5.1   2.51
6.3   8.66
7.7   1.29

then the output would be

Code:

head.dat
6   4.06
7   5.07

because 5.1 in knife.dat has already passed 5 in snake.dat so the next data point 6 would be used

and if

Code:

knife.dat
4.9   2.51
6.3   8.66
7.7   1.29

then the output would be as before

Code:

head.dat
5   3.15
6   4.06
7   5.07

because 4.9 in knife.dat has not yet passed 5 in snake.dat so a less than or equal to kind of thinking

here's kind of a picture of what the input and output should look like

Code:

snake.dat    ===========================================
knife.dat                                   ||||||||||||
head.dat                                    ============

thanks sooo much guys, tabby

tabbygirl1990 · 04-17-2014, 10:37 AM

actually, the more i play around with the example data and metaschima's command, the more i think it works, maybe your knowledge sees something i don't???

metaschima · 04-17-2014, 10:39 AM

My code seems to work with that.

Code:

bash-4.2$ awk 'FNR==NR{ if (NR == 1) knife=$1; next }{ if ($1 >= knife) print }' knife snake
6   4.06
7   5.07
bash-4.2$ awk 'FNR==NR{ if (NR == 1) knife=$1; next }{ if ($1 >= knife) print }' knife2 snake
5   3.15
6   4.06
7   5.07

grail · 04-17-2014, 10:57 AM

Here is a cut down version to do the same:

Code:

awk 'NR==1{x=$1;nextfile}$1 >= x' knife.dat snake.dat

tabbygirl1990 · 04-17-2014, 12:58 PM

thanks metaschima! it makes sense except i'm doomed to messing up FNR and NR

grail, i somewhat had something similar to yours, where you had the variable "x" i had "knifevalue" i just didn't know how to pass in something other than a user defined value. how does your do that? it starts in knife and graps the first rows first column value, then switches to the 2nd file and looks for the greater than or equal to value

grail · 04-17-2014, 01:00 PM

Quote:

it starts in knife and graps the first rows first column value, then switches to the 2nd file and looks for the greater than or equal to value

Yep

metaschima · 04-17-2014, 01:55 PM

I prefer my version as 'nextfile' is not as portable, and the print is implicit. Explicit is always better than implicit, IMO.

tabbygirl1990 · 04-17-2014, 08:32 PM

yep boys, i like both answers, muah

tabby