Delete words in a line using grep or sed - bash

I want to delete three words with a special character on a line such as
Input:
\cf4 \cb6 1749,1789 \cb3 \
Output:
1749,1789
I have tried a couple sed and grep statements but so far none have worked, mainly due to the character \.
My unsuccessful attempt:
sed -i 's/ [.\c ] //g' inputfile.ext >output file.ext

Awk accepts a regex Field Separator (in this case, comma or space):
$ awk -F'[ ,]' '$0 = $3 "." $4' <<< '\cf4 \cb6 1749,1789 \cb3 \'
1749.1789
-F'[ ,]' - Use a single character from the set space/comma as Field Separator
$0 = $3 "." $4 - If we can set the entire line $0 to Field 3 $4 followed by a literal period "." followed by Field 4 $4, do the default behavior (print entire line)
Replace <<< 'input' with file if every line of that file has the same delimeters (spaces/comma) and number of fields. If your input file is more complex than the sample you shared, please edit your question to show actual input.

The backslash is a special meta-character that confuses bash.
We treat it like any other meta-character, by escaping it, with--you guessed it--a backslash!
But first, we need to grep this pattern out of our file
grep '\\... \\... [0-9]+,[0-9]+ \\... \\' our_file # Close enough!
Now, just sed out those pesky backslashes
| sed -e 's/\\//g' # Don't forget the g, otherwise it'll only strip out 1 backlash
Now, finally, sed out the clusters of 2 alpha followed by a number and a space!
| sed -e 's/[a-z][a-z][0-9] //g'
And, finally....
grep '\\... \\... [0-9]+,[0-9]+ \\... \\' our_file | sed -e 's/\\//g' | sed -e 's/[a-z][a-z][0-9] //g'
Output:
1749,1789

My guess is you are having trouble because you have backslashes in input and can't figure out how to get backslashes into your regex. Since backslashes are escape characters to shell and regex you end up having to type four backslashes to get one into your regex.
Ben Van Camp already posted an answer that uses single quotes to make the escaping a little easier; however I shall now post an answer that simply avoids the problem altogether.
grep -o '[0-9]*,[0-9]*' | tr , .
Locks on to the comma and selects the digits on either side and outputs the number. Alternately if comma is not guaranteed we can do it this way:
egrep -o ' [0-9,]*|^[0-9,]*' | tr , . | tr -d ' '
Both of these assume there's only one usable number per line.

$ awk '{sub(/,/,".",$3); print $3}' file
1749.1789
$ sed 's/\([^ ]* \)\{2\}\([^ ]*\).*/\2/; s/,/./' file
1749.1789

Related

how to grep everything between single quotes?

I am having trouble figuring out how to grep the characters between two single quotes .
I have this in a file
version: '8.x-1.0-alpha1'
and I like to have the output like this (the version numbers can be various):
8.x-1.0-alpha1
I wrote the following but it does not work:
cat myfile.txt | grep -e 'version' | sed 's/.*\?'\(.*?\)'.*//g'
Thank you for your help.
Addition:
I used the sed command sed -n "s#version:\s*'\(.*\)'#\1#p"
I also like to remove 8.x- which I edited to sed -n "s#version:\s*'8.x-\(.*\)'#\1#p".
This command only works on linux and it does not work on MAC. How to change this command to make it works on MAC?
sed -n "s#version:\s*'8.x-\(.*\)'#\1#p"
If you just want to have that information from the file, and only that you can quickly do:
awk -F"'" '/version/{print $2}' file
Example:
$ echo "version: '8.x-1.0-alpha1'" | awk -F"'" '/version/{print $2}'
8.x-1.0-alpha1
How does this work?
An awk program is a series of pattern-action pairs, written as:
condition { action }
condition { action }
...
where condition is typically an expression and action a series of commands.
-F "'": Here we tell awk to define the field separator FS to be a <single quote> '. This means the all lines will be split in fields $1, $2, ... ,$NF and between each field there is a '. We can now reference these fields by using $1 for the first field, $2 for the second ... etc and this till $NF where NF is the total number of fields per line.
/version/{print $2}: This is the condition-action pair.
condition: /version/:: The condition reads: If a substring in the current record/line matches the regular expression /version/ then do action. Here, this is simply translated as if the current line contains a substring version
action: {print $2}:: If the previous condition is satisfied, then print the second field. In this case, the second field would be what the OP requests.
There are now several things that can be done.
Improve the condition to be /^version :/ && NF==3 which reads _If the current line starts with the substring version : and the current line has 3 fields then do action
If you only want the first occurance, you can tell the system to exit immediately after the find by updating the action to {print $2; exit}
I'd use GNU grep with pcre regexes:
grep -oP "version: '\\K.*(?=')" file
where we are looking for "version: '" and then the \K directive will forget what it just saw, leaving .*(?=') to match up to the last single quote.
Try something like this: sed -n "s#version:\s*'\(.*\)'#\1#p" myfile.txt. This avoids the redundant cat and grep by finding the "version" line and extracting the contents between the single quotes.
Explanation:
the -n flag tells sed not to print lines automatically. We then use the p command at the end of our sed pattern to explicitly print when we've found the version line.
Search for pattern: version:\s*'\(.*\)'
version:\s* Match "version:" followed by any amount of whitespace
'\(.*\)' Match a single ', then capture everything until the next '
Replace with: \1; This is the first (and only) capture group above, containing contents between single quotes.
When your only want to look at he quotes, you can use cut.
grep -e 'version' myfile.txt | cut -d "'" -f2
grep can almost do this alone:
grep -o "'.*'" file.txt
But this may also print lines you don't want to: it will print all lines with 2 single quotes (') in them. And the output still has the single quotes (') around it:
'8.x-1.0-alpha1'
But sed alone can do it properly:
sed -rn "s/^version: +'([^']+)'.*/\1/p" file.txt

Using sed to extract strings from a text file

I have text data in this form:
^Well/Well[ADV]+ADV ^John/John[N]+N ^has/have[V]+V+3sg+PRES ^a/a[ART]
^quite/quite[ADV]+ADV ^different/different[ADJ]+ADJ ^not/not[PART]
^necessarily/necessarily[ADV]+ADV ^more/more[ADV]+ADV
^elaborated/elaborate[V]+V+PPART ^theology/theology[N]+N *edu$
And I want it to be processed to this form:
Well John have a quite different not necessarily more elaborate theology
Basically, I need every string between the starting character / and the ending character [.
Here is what I tried, but I just get empty files...
#!/bin/bash
for file in probe/*.txt
do sed '///,/[/d' $file > $file.aa
mv $file.aa $file
done
awk to the rescue!
$ awk -F/ -v RS=^ -v ORS=' ' '{print $1}' file
Well John has a quite different not necessarily more elaborated theology
Explanation set record separator (RS) to ^ to separate your logical groups, also set the field separator (FS) to / and print the first field as your requirement. Finally, setting the output field separator (OFS) to space (instead of the default new line) keeps the extracted fields on the same line.
With GNU grep and Perl compatible regular expressions (-P):
$ echo $(grep -Po '(?<=/)[^[]*' infile)
Well John have a quite different not necessarily more elaborate theology
-o retains just the matches, (?<=/) is a positive look-behind ("make sure there is a /, but don't include it in the match"), and [^[]* is "a sequence of characters other than [".
grep -Po prints one match per line; by using the output of grep as arguments to echo, we convert the newlines into spaces (could also be done by piping to tr '\n' ' ').
cat file|grep -oE "\/[^\[]*\[" |sed -e 's#^/##' -e 's/\[$//' | tr -s "\n" " "

bash, text file remove all text in each line before the last space

I have a file with a format like this:
First Last UID
First Middle Last UID
Basically, some names have middle names (and sometimes more than one middle name). I just want a file that only as UIDs.
Is there a sed or awk command I can run that removes everything before the last space?
awk
Print the last field of each line using awk.
The last field is indexed using the NF variable which contains the number of fields for each line. We index it using a dollar sign, the resulting one-liner is easy.
awk '{ print $NF }' file
rs, cat & tail
Another way is to transpose the content of the file, then grab the last line and transpose again (this is fairly easy to see).
The resulting pipe is:
cat file | rs -T | tail -n1 | rs -T
cut & rev
Using cut and rev we could also achieve this goal by reversing the lines, cutting the first field and then reverse it again.
rev file | cut -d ' ' -f1 | rev
sed
Using sed we simply remove all chars until a space is found with the regex ^.* [^ ]*$. This regex means match the beginning of the line ^, followed by any sequence of chars .* and a space . The rest is a sequence of non spaces [^ ]* until the end of the line $. The sed one-liner is:
sed 's/^.* \([^ ]*\)$/\1/' file
Where we capture the last part (in between \( and \)) and sub it back in for the entire line. \1 means the first group caught, which is the last field.
Notes
As Ed Norton cleverly pointed out we could simply not catch the group and remove the former part of the regex. This can be as easily achieved as
sed 's/.* //' file
Which is remarkably less complicated and more elegant.
For more information see man sed and man awk.
Using grep:
$ grep -o '[^[:blank:]]*$' file
UID
UID
-o tells grep to print only the matching part. The regex [^[:blank:]]*$ matches the last word on the line.

Display all fields except the last

I have a file as show below
1.2.3.4.ask
sanma.nam.sam
c.d.b.test
I want to remove the last field from each line, the delimiter is . and the number of fields are not constant.
Can anybody help me with an awk or sed to find out the solution. I can't use perl here.
Both these sed and awk solutions work independent of the number of fields.
Using sed:
$ sed -r 's/(.*)\..*/\1/' file
1.2.3.4
sanma.nam
c.d.b
Note: -r is the flag for extended regexp, it could be -E so check with man sed. If your version of sed doesn't have a flag for this then just escape the brackets:
sed 's/\(.*\)\..*/\1/' file
1.2.3.4
sanma.nam
c.d.b
The sed solution is doing a greedy match up to the last . and capturing everything before it, it replaces the whole line with only the matched part (n-1 fields). Use the -i option if you want the changes to be stored back to the files.
Using awk:
$ awk 'BEGIN{FS=OFS="."}{NF--; print}' file
1.2.3.4
sanma.nam
c.d.b
The awk solution just simply prints n-1 fields, to store the changes back to the file use redirection:
$ awk 'BEGIN{FS=OFS="."}{NF--; print}' file > tmp && mv tmp file
Reverse, cut, reverse back.
rev file | cut -d. -f2- | rev >newfile
Or, replace from last dot to end with nothing:
sed 's/\.[^.]*$//' file >newfile
The regex [^.] matches one character which is not dot (or newline). You need to exclude the dot because the repetition operator * is "greedy"; it will select the leftmost, longest possible match.
With cut on the reversed string
cat youFile | rev |cut -d "." -f 2- | rev
If you want to keep the "." use below:
awk '{gsub(/[^\.]*$/,"");print}' your_file

shell replace cr\lf by comma

I have input.txt
1
2
3
4
5
I need to get such output.txt
1,2,3,4,5
How to do it?
Try this:
tr '\n' ',' < input.txt > output.txt
With sed, you could use:
sed -e 'H;${x;s/\n/,/g;s/^,//;p;};d'
The H appends the pattern space to the hold space (saving the current line in the hold space). The ${...} surrounds actions that apply to the last line only. Those actions are: x swap hold and pattern space; s/\n/,/g substitute embedded newlines with commas; s/^,// delete the leading comma (there's a newline at the start of the hold space); and p print. The d deletes the pattern space - no printing.
You could also use, therefore:
sed -n -e 'H;${x;s/\n/,/g;s/^,//;p;}'
The -n suppresses default printing so the final d is no longer needed.
This solution assumes that the CRLF line endings are the local native line ending (so you are working on DOS) and that sed will therefore generate the local native line ending in the print operation. If you have DOS-format input but want Unix-format (LF only) output, then you have to work a bit harder - but you also need to stipulate this explicitly in the question.
It worked OK for me on MacOS X 10.6.5 with the numbers 1..5, and 1..50, and 1..5000 (23,893 characters in the single line of output); I'm not sure that I'd want to push it any harder than that.
In response to #Jonathan's comment to #eumiro's answer:
tr -s '\r\n' ',' < input.txt | sed -e 's/,$/\n/' > output.txt
tr and sed used be very good but when it comes to file parsing and regex you can't beat perl
(Not sure why people think that sed and tr are closer to shell than perl... )
perl -pe 's/\n/$1,/' your_file
if you want pure shell to do it then look at string matching
${string/#substring/replacement}
Use paste command. Here is using pipes:
echo "1\n2\n3\n4\n5" | paste -s -d, /dev/stdin
Here is using a file:
echo "1\n2\n3\n4\n5" > /tmp/input.txt
paste -s -d, /tmp/input.txt
Per man pages the s concatenates all lines and d allows to define the delimiter character.
Awk versions:
awk '{printf("%s,",$0)}' input.txt
awk 'BEGIN{ORS=","} {print $0}' input.txt
Output - 1,2,3,4,5,
Since you asked for 1,2,3,4,5, as compared to 1,2,3,4,5, (note the comma after 5, most of the solutions above also include the trailing comma), here are two more versions with Awk (with wc and sed) to get rid of the last comma:
i='input.txt'; awk -v c=$(wc -l $i | cut -d' ' -f1) '{printf("%s",$0);if(NR<c){printf(",")}}' $i
awk '{printf("%s,",$0)}' input.txt | sed 's/,\s*$//'
printf "1\n2\n3" | tr '\n' ','
if you want to output that to a file just do
printf "1\n2\n3" | tr '\n' ',' > myFile
if you have the content in a file do
cat myInput.txt | tr '\n' ',' > myOutput.txt
python version:
python -c 'import sys; print(",".join(sys.stdin.read().splitlines()))'
Doesn't have the trailing comma problem (because join works that way), and splitlines splits data on native line endings (and removes them).
cat input.txt | sed -e 's|$|,|' | xargs -i echo "{}"

Resources