Python string.replace alternative

Question

I need help with Python string manipulation (I boiled down my previous looong question to this issue below).

For this line from a file:

L20B, CVS=1, HTYP=16, MLV=25

The second field could be CVS or VS. Related data will be present till the end of the line.

Need to replace the part starting with CVS or VS with another string:

if CVS found, then replacement is CFIXD(0,1,0) -OR-
if VS found, then replacement is FIXD(0,1,0)

Examples:

old line: L20B, CVS=1, HTYP=16, MLV=25
new line: L20B, CFIXD(0,1,0)

Old line: T10, M312, P10, Z3710, CL=L1, RH=1  (here, identify RH only and replace with)
New line: T10, M312, P10, Z3710, CL=L1, FIXD(0,1,0)

Old line: T20, M312, P20, Z100, CKR=10000 DV(0,1,0) 
New line: T20, M312, P20, Z100, CLS(0,1,0), MU=0.35

So, the replacement string keeps changing with what is found.
CVS or VS (till end of line) is replaced with CFIXD(0,1,0) or FIXD(0,1,0)
CRH or RH (till end of line) is replaced with CVR(0,1,0) or VR(0,1,0)
CFIXD or FIXD (till end of line) is replaced with CVR(0,1,0) or VR(0,1,0)
20 other variants.

Also, is it possible to modify the re.sub() expression to identify something in the search string and carry it over to the replacement string?
For e.g., 
Search for CFIXD(x,y,z) - replace with CVR (x,y,z)

I cannot search for the exact substring ("CVS=1, HTYP=16, MLV=25") since the data after the CVS (or VS) could be many different variations like

CVS=2, HTYP=11, MLV=25 
VS=4, HTYP=9, MLV=5      etc.

The length as you can see could be different too. Only thing I know for sure is that the string starting with CVS or VS goes till the end of that line. As far as I know, string.replace won't work because of the above varying length and data.

Any readily available Python methods? Or do I have to write a small routine to do this? I can find the index (using string.find) to the VS or CVS and then replace everything from that point to the end of the line, yes? I know there is a simple (not for me) regex way. Thanks.

Am I correct in understanding that you wish to grab everything after and including the characters "CVS=" or "VS="? It is not clear what needs to be replaced. Can you give example input and output strings? This should be a simple regex exercise. — Cory Dolphin
– Cory Dolphin, Commented Dec 13, 2012 at 19:52

Andrew Clark · Accepted Answer · 2012-12-13 20:59:43Z

3

Use regular expressions:

import re
re.sub(r'(C|)VS=.*', r'\1FIXD(0,1,0)', line)

Explanation:

# regex
  (C|)         # optionally match a 'C', save it or an empty string in group 1
  VS=          # match 'VS='
  .*           # match to the end of the line

# replacement
  \1           # the contents of group 1 (either 'C' or an empty string)
  FIXD(0,1,0)  # the literal string 'FIXD(0,1,0)'

Examples:

>>> re.sub(r'(C|)VS=.*', r'\1FIXD(0,1,0)', 'L20B, CVS=1, HTYP=16, MLV=25')
'L20B, CFIXD(0,1,0)'
>>> re.sub(r'(C|)VS=.*', r'\1FIXD(0,1,0)', 'L20C, VS=4, HTYP=9, MLV=5')
'L20C, FIXD(0,1,0)'

Edit: based on your edit here are a few alternatives for your different cases:

CVS or VS -> CFIXD(0,1,0) or FIXD(0,1,0)

re.sub(r'(C|)VS=.*', r'\1FIXD(0,1,0)', line)

CRH or RH -> CVR(0,1,0) or VR(0,1,0)

re.sub(r'(C|)RH=.*', r'\1VR(0,1,0)', line)

CFIXD(x,y,z) or VIXD(x,y,z) -> CVR(x,y,z) or VR(x,y,z)
```
re.sub(r'(C|)FIXD(\([^)]*\)).*', r'\1VR\2', line)
```

Explanation of (\([^)]\)).*:

(         # start second capture group
   \(       # match a literal '('
   [^)]*    # match any number of characters that are not ')'
   \)       # match a literal ')'
)         # end capture group
.*        # match to the end of the line

edited Dec 13, 2012 at 20:59

answered Dec 13, 2012 at 19:53

Andrew Clark

210k36 gold badges285 silver badges310 bronze badges

Sign up to request clarification or add additional context in comments.

11 Comments

user1858112 Over a year ago

Thanks, that looks simple. But, will it add the 'C' if it finds it?

Andrew Clark Over a year ago

Yes, but it is probably better to use a solution that you understand. If you have specific questions about how this works feel free to ask.

user1858112 Over a year ago

Actually, the last field (replacement) needs to be one of many choices. CVS/VS replace with FIXD() or CFIXD, RH/CRH replace with CLS()...... or LS()......., and so on. But, if I figure out the mechanism for one, I can handle the others.

Andrew Clark Over a year ago

Ahh got it, if you are having trouble adapting it to your other fields, you should be able to use re.sub(r'(C|)RH=.*', r'\1LS(0,1,0)', line) for the RH/CRH replacements.

user1858112 Over a year ago

Sorry, I just elaborated with more examples. Thank you kindly for assisting me.

|

kindall · Accepted Answer · 2012-12-13 20:03:21Z

1

You can use a regular expression, but to me it'd be more straightforward to just split the string into a list:

line = "L20B, CVS=1, HTYP=16, MLV=25"

line = line.split(", ")
if line[1].startswith("CVS="):
    line[1:] = ["CFIXD(0,1,0)"]
elif line[1].startswith("VS="):
    line[1:] = ["FIXD(0,1,0)"]

line = ", ".join(line)

These two cases could be combined with some jiggery-pokery, since they are rather similar, but they seem perfectly readable this way.

answered Dec 13, 2012 at 20:03

kindall

185k36 gold badges291 silver badges321 bronze badges

3 Comments

user1858112 Over a year ago

I agree but the problem is that VS could show up towards the end of the line. Example: T10,M312,P10,Z3710,CL=L1, VS=1,HTYP=16,MLV=25; in this case, I guess I could modify with a loop, say line[i].startswith("VS="). Is that right?

kindall Over a year ago

Yes, although you're starting to get into "regex is easier" territory here.

user1858112 Over a year ago

I like your solution but it gets verbose quickly with the 20 variants I have. But, it is understandable to a non-regex person like me. I will see how it goes. Thank you both for your quick help.

Collectives™ on Stack Overflow

Python string.replace alternative

2 Answers 2

11 Comments

3 Comments

Your Answer

Linked

Hot Network Questions

Collectives™ on Stack Overflow

2 Answers 2

11 Comments

3 Comments

Your Answer

Sign up or log in

Post as a guest

Linked

Related