Any suggestions to improve Python string parsing-CodePudding

I'm running Python 3.6.8. I need to sum values that appear in a log file. The line may contain 1 to 14 {index,value} pairs; a typical line for 8 values is in the code below(variable called 'log_line'). The line format with the '- -' separator is consistent. I have working code, but I'm not sure if this is the most elegant or best way to parse this string; it feels a bit clunky. Any suggestions?

    import re
    
    #verion 1
    log_line = 'Some explanatory text was here:      - -{0, 8} {1, 24} {2, 24} {3, 5} {4, 5} {5, 12} {6, 12} {7, 5}'
    log_line_values = log_line.split('- -')[1]
    values = re.findall(r'{\d ,\s\d }',log_line_values)
    sum_of_values = 0
    for v in values:
        sum_of_values  = int(v.replace('{','').replace('}','').replace(' ','').split(',')[1])
    print(f'1) sum_of_values:{sum_of_values}')

    #verions 2, essentially the same, but more concise (some may say confusing)
    sum_of_values = sum([int(v.replace('{','').replace('}','').replace(' ','').split(',')[1]) for v in re.findall(r'{\d ,\s\d }',log_line.split('- -')[1])])
    print(f'2) sum_of_values:{sum_of_values}')

CodePudding user response：

First, no need to get rid of the prefix - the regex will take care of not matching that. Second, we can use capturing groups to capture values that we only care about. In our case, the second value in a comma seperated pair. We can use map(int, iterable) to turn every string to an int in a list, and then we can use sum on that list of numbers.

Putting it all together:

import re

log_line = 'Some explanatory text was here:      - -{0, 8} {1, 24} {2, 24} {3, 5} {4, 5} {5, 12} {6, 12} {7, 5}'
values = re.findall(r'{\d ,\s(\d )}', log_line_values)
sum_of_values = sum(map(int, values))

CodePudding user response：

Ideal use case for regular expressions capture groups:

import re

log_line = 'Some explanatory text was here:      - -{0, 8} {1, 24} {2, 24} {3, 5} {4, 5} {5, 12} {6, 12} {7, 5}'
pattern = r'{(\d ), (\d )}'

s = sum([int(e[1]) for e in re.findall(pattern, log_line.split('- -')[1])])

print(s) # 95

Here I use re.findall to match numbers from input array and use list comprehension to convert them to numbers and sum.

The advantage of using {(\d ), (\d )} pattern is the ability to extract first number too (if you need it).

CodePudding user response：

Assuming you've already identified that the line is one that matches the pattern, you can simplify your logic a lot by using a generator expression within sum().

import re

# Compile your regular expression for reuse
# Just pull out the last value from each pair
re_extract_val = re.compile(r'{\d , (\d )}')

log_line = 'Some explanatory text was here:      - -{0, 8} {1, 24} {2, 24} {3, 5} {4, 5} {5, 12} {6, 12} {7, 5}'

# Use generator comprehension within sum() to add all values
sum_of_values = sum(int(val) for val in re_extract_val.findall(log_line))

You could also use map(), but I find it's clearer with a generator expression

sum_of_values = sum(map(int, re_extract_val.findall(log_line)))