Shell Programming and Scripting

Hi expert

The delete the 2th

Code:

lyang0@lyang0-OptiPlex-755:~/del$ sed 's/[a-zA-Z]*//2p' 9

to delete from 2th

Code:

lyang0@lyang0-OptiPlex-755:~/del$ sed 's/[a-zA-Z]*//2gp' 9

but if I want to delete 2th 3rd 4th fileds

or

but if I want to delete 2th 15th fileds

how could I do?

Try like...

Code:

 sed '1~4d' test1.txt

Your request is extremely vague. What do you mean by?:
Do you want to delete the 2nd, 3rd, and 4th lines in the file?
Do you want to delete the 2nd, 3rd, and 4th fields in each line in the file?
What is in the lines other than unaccented uppercase letters and unaccented lowercase letters? If the fields are separated from each other by one or more space or tab characters do you also want the space and tab characters following the fields that are being removed to be removed or do you want to leave the separators unchanged?
Do you want to only print lines that have been changed, or do you want to every line after the changes have been completed. (Your current code appears to want to print unchanged lines once and changed lines twice.)

According to the standard, the use of 2g in

Code:

s/[a-zA-Z]*//2gp

produces undefined results.
You can specify a count or the g flag, but not both.
Since you don't show us any of your input, you don't give us the output from the commands you listed, and you don't show us the output you want to get from a representative input file, any suggestions you get here will be wild conjecture as to whether or not you will get anything resembling what you want.

With the sed commands you are showing, I assume that you are trying to remove fields rather than lines, that you want to leave any field separators unchanged, and that you want to every line in the file once. In that case, something like:

Code:

sed 'sed 's/[a-zA-Z][a-zA-Z]*//14
s/[a-zA-Z][a-zA-Z]*//4
s/[a-zA-Z][a-zA-Z]*//3
s/[a-zA-Z][a-zA-Z]*//2' 9

will delete two or more letters in the 2nd, 3rd, 4th, and 14th nonadjacent occurrences of adjacent letters from every line. If a line has fewer occurrences of this condition, it will still make changes to any of the times the request is matched. If the input file (9) contains:

Code:

first second third fourth fifth sixth seventh eighth ninth tenth eleventh twelfth thirteenth fourteenth fifteenth sixteenth seventeenth eighteenth nineteenth twentieth
1st 2nd 3rd 4th 5th 6th 7th 8th 9th 10th

the output will be:

Code:

first    fifth sixth seventh eighth ninth tenth eleventh twelfth thirteenth  fifteenth sixteenth seventeenth eighteenth nineteenth twentieth
1st 2 3 4 5th 6th 7th 8th 9th 10th

can we use one line sed to handle this ?

---------- Post updated at 07:35 AM ---------- Previous update was at 07:32 AM ----------


I want to delete the 2nd, 3rd, and 4th fields in each line in the file, can we use one line sed to handle this?

To delete the 2nd, 3rd, and 4th fields (using the same assumptions I made above about what you define a field to be), the following single line sed will handle it:

Code:

sed 's/[a-zA-Z][a-zA-Z]*//4;s/[a-zA-Z][a-zA-Z]*//3;s/[a-zA-Z][a-zA-Z]*//2' 9

it is just harder to read than the same code with the <semicolon>s replaced by <newline>s. It can't be done with a single sed substitute command. (At least it can't in a portable way. Some implementations of sed might support providing a list and/or range of values when specifying which occurrence of a match for the BRE in a substitute command, but I haven't seen any that do so.

The following one sed command seems portable:

Code:

sed 's/\([a-zA-Z]*\) *[a-zA-Z]* *[a-zA-Z]* *[a-zA-Z]*/\1/'

The person who started this thread still hasn't told us what the input format is. This one line will remove the 2nd, 3rd, and 4th fields if and only if there are no characters other than spaces and unaccented alphabetic characters in the before the end of the 4th string of letters ends.

The submitter also asked how to delete the 2nd and 15th fields. That can still be done as a single replacement but it is tedious to type and debug at best.

We need a CLEAR statement of what a field is from the submitter before any of us will know if what we might suggest will work.

My script will take the input file:

Code:

1st 2nd 3rd{tab}4th 5th

where {tab} is a tab character and produce:

Code:

1st 2 3{tab}4 5th

With the same input, your script will just copy the input to the output without change. I have no idea which if either of these is what the submitter wants since there is no definition of what is in the file other than some alphabetic characters.