# Help required on AWK command **URL:** \ **Category:** UNIX for Dummies Questions & Answers\ **Created:** [January 21, 2008, 3:10am UTC](https://community.unix.com/t/help-required-on-awk-command/182478 "2008-01-21T03:10:04Z")\ **Posts on this page:** 14\ **Page:** 1
**Author:** ![udiptya](https://community.unix.com/letter_avatar/udiptya/32/6_449c533f586939b6e5f246632a46421f.png) [@udiptya](https://community.unix.com/u/udiptya)\ **Post date:** [January 21, 2008, 3:10am UTC](https://community.unix.com/t/help-required-on-awk-command/182478/1 "2008-01-21T03:10:04Z")
I have two questions: (Q1) I am using a loop and want to print a particular field of a table using the AWK command by accessing one row at a time by validating NR with the loop counter. I am using a command like this but it isn't working! count=1 NAWK -F"|" -v var=$count 'NR == var { printf $5 }' file1 where, count is my loop counter variable; file1 is the table to search. I will appreciate if someone gives me a solution (Q2) I am taking a name as input and storing it in a variable empname Now, the situation is same as in question 1 and I have used a command: NAWK -F"|" -v var=$count 'NR = var { printf $5 }' file1 in this case although I am using the variable count=1 for printing field 5 of the first row the answer is given according to the variable empname as if I have used a command something like NAWK -F"|" -v var="${empname}" '$1 ~ var { printf $5 }' file1 I will appreciate if someone gives me an explanation of this ---
**Author:** ![ThobiasVakayil](https://community.unix.com/letter_avatar/thobiasvakayil/32/6_449c533f586939b6e5f246632a46421f.png) [@ThobiasVakayil](https://community.unix.com/u/ThobiasVakayil)\ **Post date:** [January 21, 2008, 4:00am UTC](https://community.unix.com/t/help-required-on-awk-command/182478/2 "2008-01-21T04:00:18Z")
count=1 awk -F"|" '$1 ~/'$count'/ {print $5}' file1 ---
**Author:** ![udiptya](https://community.unix.com/letter_avatar/udiptya/32/6_449c533f586939b6e5f246632a46421f.png) [@udiptya](https://community.unix.com/u/udiptya)\ **Post date:** [January 21, 2008, 11:57pm UTC](https://community.unix.com/t/help-required-on-awk-command/182478/3 "2008-01-21T23:57:55Z")
Hi, Thank you for your initiative but the suggestion isn't working. Can you suggest something else. ---
**Author:** ![ThobiasVakayil](https://community.unix.com/letter_avatar/thobiasvakayil/32/6_449c533f586939b6e5f246632a46421f.png) [@ThobiasVakayil](https://community.unix.com/u/ThobiasVakayil)\ **Post date:** [January 22, 2008, 12:04am UTC](https://community.unix.com/t/help-required-on-awk-command/182478/4 "2008-01-22T00:04:53Z")
please check the following : cat file1 1|aaa|bbb|ccc 2|xxx|yyy|zzz 3|mmm|nnn|ooo $ count=2 $ awk -F"|" '$1 ~ /'$count'/ {print $2}' file1 output : xxx ---
**Author:** ![udiptya](https://community.unix.com/letter_avatar/udiptya/32/6_449c533f586939b6e5f246632a46421f.png) [@udiptya](https://community.unix.com/u/udiptya)\ **Post date:** [January 22, 2008, 2:33am UTC](https://community.unix.com/t/help-required-on-awk-command/182478/5 "2008-01-22T02:33:20Z")
Sorry, I should have given a bit of detail. My file is like this CAT file1 abc |123 |abc123 bcd |234 |bcd234 cde |345 |cde345 In such a situation the command will not work. I want to extract one field from this file by scrolling through its individual lines which I want to do within a loop. I also would like to ask something on your feedback too: Now considering I upgrade my file as you suggested like CAT file1 1 |abc |123 |abc123 2 |bcd |234 |bcd234 3 |cde |345 |cde345 . . 100 |uvw |999 |uvw999 In this situation also your command will give erronous results. What should I do in such a situation? ---
**Author:** ![udiptya](https://community.unix.com/letter_avatar/udiptya/32/6_449c533f586939b6e5f246632a46421f.png) [@udiptya](https://community.unix.com/u/udiptya)\ **Post date:** [January 22, 2008, 2:42am UTC](https://community.unix.com/t/help-required-on-awk-command/182478/6 "2008-01-22T02:42:43Z")
I also tried with a slight modification of your command AWK -F"|" '$1 == /'$count'/ {prinf $4}' file1 but this didn't work either ---
**Author:** ![ThobiasVakayil](https://community.unix.com/letter_avatar/thobiasvakayil/32/6_449c533f586939b6e5f246632a46421f.png) [@ThobiasVakayil](https://community.unix.com/u/ThobiasVakayil)\ **Post date:** [January 22, 2008, 3:25am UTC](https://community.unix.com/t/help-required-on-awk-command/182478/7 "2008-01-22T03:25:37Z")
In my case the following command is working fine : $ count=2 $ awk -F"|" '$1 ~ /'$count'/ {prinf $4}' file1 The below mentioned command won't work : awk -F"|" '$1 == /'$count'/ {prinf $4}' file1 == won't work. Please try the first command and send the error details. ---
**Author:** ![udiptya](https://community.unix.com/letter_avatar/udiptya/32/6_449c533f586939b6e5f246632a46421f.png) [@udiptya](https://community.unix.com/u/udiptya)\ **Post date:** [January 22, 2008, 3:44am UTC](https://community.unix.com/t/help-required-on-awk-command/182478/8 "2008-01-22T03:44:28Z")
Your command is working fine for the type of file you gave as an example. But consider the file I have given as an example in such a case the command count=1 AWK -F"|" '$1 ~ /'$count'/ {printf $4}' file1 will give the output for all the numbers starting with 1 (eg 1, 100, 101..etc)in the first column of the table! Since ~ gives an approimate matching this won't help. I need an exact watch which is possible with == ---
**Author:** ![ThobiasVakayil](https://community.unix.com/letter_avatar/thobiasvakayil/32/6_449c533f586939b6e5f246632a46421f.png) [@ThobiasVakayil](https://community.unix.com/u/ThobiasVakayil)\ **Post date:** [January 22, 2008, 3:55am UTC](https://community.unix.com/t/help-required-on-awk-command/182478/9 "2008-01-22T03:55:11Z")
Hello, If you want to get the line starts with 100 means, you have to give like this : count=100 awk -F"|" '$1 ~ /'$count'/ {print $3} file1 you can also give the search string like this : $ search\_string="aaa" $ awk -F"|" '$2 ~ /'$search\_string'/ {print $3} file1 count is not the line number. It's a search string only. Hope you got better idea. ---
**Author:** ![udiptya](https://community.unix.com/letter_avatar/udiptya/32/6_449c533f586939b6e5f246632a46421f.png) [@udiptya](https://community.unix.com/u/udiptya)\ **Post date:** [January 22, 2008, 4:59am UTC](https://community.unix.com/t/help-required-on-awk-command/182478/10 "2008-01-22T04:59:07Z")
I have understood your point of view btu I guess I couldn't properly express you with my problem. So, I am writing a part of my program underneath my file is something like this with tabs and whitespaces inbetween and delimited by "|' CAT file1 abc |123 |234 bcd |345 |456 cde |567 |678 Now my program is something like this: start\_cnt=1 end\_cnt=200 while [start\_cnt -le end\_cnt] do var1=`AWK -F"|" '...........{ printf $2 }' file1` echo "$var1" \>\> file2 start\_cnt=`expr $start_cnt + 1` done echo "Finished file appending" Now I don't know what to write in the space provided in the AWK command such that it will use the start\_cnt as the incrementing variable starting with line 1 as start\_cnt=1 and line 200 as start\_cnt=200 and on each step will append the value at position $2 to file2 I hope now I could make you understand my problem! ---
**Author:** ![ThobiasVakayil](https://community.unix.com/letter_avatar/thobiasvakayil/32/6_449c533f586939b6e5f246632a46421f.png) [@ThobiasVakayil](https://community.unix.com/u/ThobiasVakayil)\ **Post date:** [January 22, 2008, 5:19am UTC](https://community.unix.com/t/help-required-on-awk-command/182478/11 "2008-01-22T05:19:12Z")
I think no need to use the loop in here. You can just use the awk command and write the output to a file like this : search\_string="xxx" awk -F"|" '$1 ~ '/$search\_string/' { printf $2 }' file1 \> file2 ---
**Author:** ![udiptya](https://community.unix.com/letter_avatar/udiptya/32/6_449c533f586939b6e5f246632a46421f.png) [@udiptya](https://community.unix.com/u/udiptya)\ **Post date:** [January 22, 2008, 6:14am UTC](https://community.unix.com/t/help-required-on-awk-command/182478/12 "2008-01-22T06:14:40Z")
Yes, that can be done but for a long table if you have to search each element of the field, you will require a lot of code. Whereas I am looking for a very genaralised pattern! ---
**Author:** ![vgersh99](https://community.unix.com/user_avatar/community.unix.com/vgersh99/32/14851_2.png) [@vgersh99](https://community.unix.com/u/vgersh99)\ **Post date:** [January 22, 2008, 6:41am UTC](https://community.unix.com/t/help-required-on-awk-command/182478/13 "2008-01-22T06:41:03Z")
It's a bit expensive to scan the ENTIRE input file on EVERY iteration of the loop, but... Given udi.txt: ```nohighlight abc |123 |234 bcd |345 |456 cde |567 |678 ``` ```nohighlight #!/bin/sh start_cnt=1 end_cnt=200 while ["${start_cnt}" -le "${end_cnt}"] do var1=`awk -F"|" -v fnr="${start_cnt}" 'FNR == fnr { printf $2 }' udi.txt` echo "$var1" start_cnt=`expr $start_cnt + 1` done echo "Finished file appending" ``` ---
**Author:** ![udiptya](https://community.unix.com/letter_avatar/udiptya/32/6_449c533f586939b6e5f246632a46421f.png) [@udiptya](https://community.unix.com/u/udiptya)\ **Post date:** [January 22, 2008, 7:43am UTC](https://community.unix.com/t/help-required-on-awk-command/182478/14 "2008-01-22T07:43:03Z")
Thanks a lot. It worked! I tried that way but I was using NR in place FNR. Thanks again!