How To Find All Repeated Pattern In One String Using Regular Expression

Question

[cauyrd]

I have a following string:

'0\t36aa, >HWUSI-EAS614_4:1:62:3987:7089:0:1:1... *1\t36aa, >HWUSI-EAS614_4:1:65:3993:10262:0:1:1... at 100.00%2\t36aa, >ILLUMINA-EAS295_5:3:99:4680:15673:0:1:1... at 100.00%3\t36aa, >HWUSI-EAS614_4:1:63:11191:7359:0:1:1... at 94.44%`

I wanna find all pattern between '>' and '...', which are:

HWUSI-EAS614_4:1:62:3987:7089:0:1:1
HWUSI-EAS614_4:1:65:3993:10262:0:1:1
ILLUMINA-EAS295_5:3:99:4680:15673:0:1:1
HWUSI-EAS614_4:1:63:11191:7359:0:1:1

How to write the regular expression, python syntax is preferred.

Answer 1

[JC]

perl -lane 'if (m/>(.+?)\.\.\./) { print $1; }' < FILE

edit: ok, no cat

edit2: the original question had one sequence ID per line, now it's showed with all IDs in one line, to satisface the comments below:

perl -lane 'print $1 while (m/>(.+?)\.\.\./g)' < FILE

[Ben]

UUOC

[SES]

Useless use of redirection also, plus it doesn't do what the OP requested. The answer from @qiyunzhu below will print all the matches.

[JC]

when I wrote my answer, the question had the string with one ID per line, not all in one line

[SES]

Okay. It's hard to tell when there have been edits, but that first solution would work then.

Answer 2

[Sukhi Singh]

Use cut with the flexible -d parameter.

I assume all of the lines start with > and ends with ... and there won't be a period in between, then

cut -f2 -d'>' FILE | cut -f1 -d'.'

gives you what you want.

Another one using sed ;

sed -e 's/.*>\(.*\)\.\.\..*/\1/g' FILE

Answer 3

[qiyunzhu]

Here's a Perl solution. The trick is to remove one match per time.

print $1 while s/\>(.+?)\.\.\.//;

Answer 4

[Whetting]

since python syntax was requested...

import re 

string="0\t36aa, >HWUSI-EAS614_4:1:62:3987:7089:0:1:1...
*1\t36aa, >HWUSI-EAS614_4:1:65:3993:10262:0:1:1... at 100.00%2\t36aa, >ILLUMINA-EAS295_5:3:99:4680:15673:0:1:1... at 100.00%3\t36aa, >HWUSI-EAS614_4:1:63:11191:7359:0:1:1... at 94.44%"  


print re.findall(">(.+?)\.\.\.",string)

Answer 5

[Manu Prestat]

import re    #the regexp python module

myPattern = re.compile('\>(.+?)\.{3}')
myText ='0\t36aa, >HWUSI-EAS614_4:1:62:3987:7089:0:1:1... *1\t36aa, >HWUSI-EAS614_4:1:65:3993:10262:0:1:1... at 100.00%2\t36aa, >ILLUMINA-EAS295_5:3:99:4680:15673:0:1:1... at 100.00%3\t36aa, >HWUSI-EAS614_4:1:63:11191:7359:0:1:1... at 94.44%'
myPatRes = myPattern.findall(myText)
print myPatRes

['>HWUSI-EAS6144:1:62:3987:7089:0:1:1...', '>HWUSI-EAS6144:1:65:3993:10262:0:1:1...', '>ILLUMINA-EAS2955:3:99:4680:15673:0:1:1...', '>HWUSI-EAS6144:1:63:11191:7359:0:1:1...']

I guess in your case, you want to simplify:

for res in myPatRes:
    print res[1:-3]

HWUSI-EAS614_4:1:62:3987:7089:0:1:1

HWUSI-EAS614_4:1:65:3993:10262:0:1:1

ILLUMINA-EAS295_5:3:99:4680:15673:0:1:1

HWUSI-EAS614_4:1:63:11191:7359:0:1:1

[Manu Prestat]

I realize that while I was writting it, Whetting answers the short (and good!) version. Here is the longer story ;-)