[SOLVED] Delete from second part of each line and go to next line

sam@ · 02-13-2016, 12:24 AM

Hi ,
My input file looks like this:

Code:

>AHMI01000729.1 Details of items list product_1, whole details set here
>AHMI01000113.1 Details of items list product_10, whole details set here
>AHMI01000012.1 Details of items list product_100, whole details set here
>AHMI01002223.1 Details of items list product_9953, whole details set here
>AHMI01002226.1 Details of items list product_997, whole details set here
>AHMI01002225.1 Details of items list product_9971, whole details set here

My required output is as follows:

Code:

>item729
>item113
>item12
>item2223
>item2226
>item2225

I tried following sed script:

Code:

sed 's/AHMI01* Details of items list product _*, whole details set here/item/g'  inputfile> outputfile

but it didnt work out!

What could I possibly change?

Thanks!

Michael Uplawski · 02-13-2016, 01:37 AM

IGNORE. All wrong. I ignored that you skip the zeros.

Do not laugh at me, but I rather use a programming language than plunge into sed. How about

Code:

for i in `cat in.txt | cut -d' ' -f1|cut -c6-13`; do echo ">item$i";done

... and be it for the record.

sycamorex · 02-13-2016, 02:25 AM

Quote:

Originally Posted by sam@

Hi ,
My input file looks like this:

Code:

>AHMI01000729.1 Details of items list product_1, whole details set here
>AHMI01000113.1 Details of items list product_10, whole details set here
>AHMI01000012.1 Details of items list product_100, whole details set here
>AHMI01002223.1 Details of items list product_9953, whole details set here
>AHMI01002226.1 Details of items list product_997, whole details set here
>AHMI01002225.1 Details of items list product_9971, whole details set here

My required output is as follows:

Code:

>item729
>item113
>item12
>item2223
>item2226
>item2225

I tried following sed script:

Code:

sed 's/AHMI01* Details of items list product _*, whole details set here/item/g'  inputfile> outputfile

but it didnt work out!

What could I possibly change?

Thanks!

Hi there. Here you go:

Code:

sed 's/\([^1]*\)\(1[^[1-9]*\)\([^\.]*\).*/>item\3/' file

Code:

sycamorex@darkstar:~/data/tmp$ cat file1
>AHMI01000729.1 Details of items list product_1, whole details set here
>AHMI01000113.1 Details of items list product_10, whole details set here
>AHMI01000012.1 Details of items list product_100, whole details set here
>AHMI01002223.1 Details of items list product_9953, whole details set here
>AHMI01002226.1 Details of items list product_997, whole details set here
>AHMI01002225.1 Details of items list product_9971, whole details set here
sycamorex@darkstar:~/data/tmp$ sed 's/\([^1]*\)\(1[^[1-9]*\)\([^\.]*\).*/>item\3/' file1
>item729
>item113
>item12
>item2223
>item2226
>item2225

grail · 02-13-2016, 04:21 AM

The sed can reduce the back references a little:

Code:

sed -r 's/[^1]*10*([^.]*).*/>item\1/' file

And awk works too:

Code:

awk -F"[0.]+" '{print ">item"$3}' file

A.Thyssen · 02-14-2016, 07:30 PM

Quote:

Originally Posted by grail

And awk works too:

Code:

awk -F"[0.]+" '{print ">item"$3}' file

Awk code will fail if there is a zero in the item number!

For example (made up based on previous examples)

Quote:

>AHMI01000709.1 Details of items list product_53243, whole details set here

Would result in

Quote:

item7

which is obviously wrong.

try this, which based on position of the number within the string as it seems constant

Code:

awk '{print ">item" substr($0,8,6)+0}'