UNIX for Dummies Questions & Answers

Hi,


After looking at the differents post on this forum, I am convinced that I will benefit from the experience of advanced Unix user on some script I have already done for an aeronautical study. Here is one of them :

Step 1 :

Code:

sed -e "s/??/00/g" Base_Awk.txt > Awk_Cut_00.txt4;
sed -e "s/??/01/g" Base_Awk.txt > Awk_Cut_01.txt4;
sed -e "s/??/02/g" Base_Awk.txt > Awk_Cut_02.txt4;
cat *.txt4 > Awk_Cut_All.sh;rm *txt4;

with Base_Awk.txt :

Code:

awk 'NR==1{system("cat Cut_1??.txt")}1' Fichier_Base.txt >   Fichier_Base_1.txt;
awk 'NR==5001{system("cat Cut_2??.txt")}1' Fichier_Base_1.txt > Fichier_Base_2.txt;
...
awk 'NR==30001{system("cat Cut_7??.txt")}1' Fichier_Base_6.txt > Statique_6oc_Cut_??.txt;

Step 2

Code:

./Awk_Cut_All.sh

To summarize, I have 7 x 70 files Cut_1??.txt,Cut_2??.txt,...,Cut_7??.txt and the output is 70 files ordered with the awk command (line 1 file Cut_101.txt line 5001 Cut_2??.txt and so on to Cut_7??.txt line 30001).

I have obtained this results with the code above, I was wondering if i could bypass the script with the sed replacement and using variables in the awk command, something like that (I know it won(t work just for the idea) :

Code:

i=[1-70];
for i in [1-70];
do 
awk 'NR==1{system("cat Cut_1$I.txt")}1' Fichier_Base.txt > Fichier_Base_1$i.txt
...
awk 'NR==1{system("cat Cut_7$I.txt")}1' Fichier_Base.txt > Fichier_Base_7$i.txt
cat Fichier_Base_?$i.txt > output_$i.txt5;

I have seen on this forum the use of variable in Awk but have been unable to use it for this script.

If anyone can lend me a hand on this one, I would be much grateful.

Marc

The easiest way to use shell variables in an awk script is to pass them in:

Code:

export SHELLVAR="Pie"; echo | awk -vAWKVAR="$SHELLVAR" '{print AWKVAR}'

Why don't you run the loop within awk to be more efficient?

I'm sorry I don't absolutely understand what you are heading for, but I'd bet there would be more efficient ways to achieve your goals. We could help if you'd post sample input and desired output files and the logics that connect them.

I have attached the maximum of 5 files per post but it will be enough I hope.

The Cut_?0?.txt files are the input, as output for this example let's say I want the Desired_Output_Cut_01(Cut_101_Cut_201).txt file and another which would correspond to Cut_02 with Cut_102.txt and Cut_202.txt at line 50 concatenated.

In fact, the first digit match an envelop of loadcase (?01), the last two digit correspond to a cut to obtain the flux in the area (1??). On the real study, I have obtained the correct output with the scripts enclosed in the first post but as RudiC pointed out a loop like the one described in the end of my first post is the code I would like to see working. It might spare me a lot of time. I have not been able to use variable in the loop (Thanks CarloM for the syntax).

Sorry, still not clear. Your Desired_Output_Cut_01(Cut_101_Cut_201).txt is just a concatenation of two of the input files. You don't need awk for that.
If I read your approaches in post#1 correctly, you have a file and want the cut- files inserted every 5000 lines? How's all that connected to Base_Awk.txt and Fichier_Base.txt?

Re-reading myself, I reckon I have been difficult to understand.
As Input, let's say I have 90 files named Cut_£??.txt (with ?? matching the cut from 01 ->10 and £ matching the envelop from 1 ->9).

I want as output; 10 files one per cut and in each one a cat of the different envelop with the awk command to have the different envelop at the correct line.

I did that with the sed script and the Base_Awk.txt of Step 1, the Fichier_Base.txt file is an empty file with 50 000 line because I did not manage to insert a file at line 25000 for example when the input file was not 25000 line long.

If you are using bash, try this (set -vx options to see what happens):

Code:

for i in {1..9}; do cat cut_${i}{01..10}.txt > output_$i.txt; done