使用startswith和index从结构化字符串中查找substring

Question

I am trying to create a code that finds a substring (can be a number or anything else) from a structured string. The string is structured (2 possibilites) like:

1. string = '1x substring 3x 4x'

2. string = '4x 3x substring 1x'

   • x can be any character
   • substring formatted like 'pos. 2' 'pos. 2'

The normal case works with the code below, but now I would also like to consider the special cases.I have tried: i.startswith(('3','4')) , but that didn't work.

string 1-8 should explain the logic using a simple example.

string 9-10 shows a complex example. The code should extract the substrings at Pos 2, 5 and Pos 7.

I hope you can help to find a solution for all strings / special cases to get clean: 80 for all cases. :-)

string 9:

clean: pos2 ='$80' pos5 = '75.000 kg' pos7 = '22 sec'

string 10:

clean: pos2 = '$67' pos5 = '69.000kg' pos7 = '12sec'

#str1-8 easy example strings

#normal string
str1 = '1x 80 3x 4x'
str2 = '4x 3x 80 1x'

# missing number/pos. 3
str3 = '1x 80 4x '
str4 = '4x 80 1x'
str3a = '1. A  $67  4. A  69.000kg  6. A  12sec  8. B  9. B'
result: `clean: pos2 = '$67'   pos5 = '69.000kg'    pos7 = '12sec'` 
str4a = '9B 8B 22 sec 6A 75.000kg 4b  $80 1b'
result: `clean:  pos2 ='$80'    pos5 = '75.000 kg'   pos7 = '22 sec'` 

# missing number/pos. 1, => number is at the start or end of the string
str5 = '80 3x 4x'
str6 = '4x 3x 80'
str5a =  '10 Mrd 3: A 4: A  50 .379 6: A   7:19   8: B 9: D ' 
result: clean: pos2= 10 Mrd, pos5= 50,379 pos7=7:19 (or just 19 in raw string without 7: if its easier)
str6a = '  9a 8b 10 6b 60000 4a 3 b 50 '
result: clean: pos2= 50, pos5= 60000 pos7=10 

# Optional (rare case)
# missing number/pos. 1 and 3
str7 = '80 4x'
str8 = '4x 80'
str7a = '10 Mrd 4: A  50 .379 6: A   7:19   8: B 9: D ' 
result: clean: pos2= 10 Mrd, pos5= 50,379 pos7=7:19 (or just 19 in raw string without 7: if its easier)
str8a = ' 9a 8b 10 6b 60000 4a  50 '
result: clean: pos2= 50, pos5= 60000 pos7=10 
# complex realistic strings
str9 = '9B 8B 22 sec 6A 75.000kg 4b 3b  $80 1b'
str10 = '1. A  $67  3. A  4. A  69.000kg  6. A  12sec  8. B  9. B'

# missing number/pos. 4 or 6 (Pos6 Optional, cause thats difficult i guess)
str11 = '1. A  $67  3. A    69.000kg  6. A  12sec  8. B  9. B'
result: `clean: pos2 = '$67'   pos5 = '69.000kg'    pos7 = '12sec'
str12 = '1. A  $67  3. A  4a  69.000kg   12sec  8. B  9. B'
result: `clean: pos2 = '$67'   pos5 = '69.000kg'    pos7 = '12sec'

x_list = [str1,str2,str3,str4,str5,str6,str7,str8, str9, str10, str11,str12]

for x in x_list:
    print ("raw             "+x)
    
    values = ['1x', '3x', '4x']
    try:
        for i in values:
            if i.startswith('3') :
                foo=i

            if  i.startswith("1") :
                baa=i 

            start=x.index(foo) + len( foo )
            end=x.index(baa)   

            if start < end:
                pass
                number = x[start:end].strip(' ')


            else:
                start=x.index(baa) + len( baa )
                end=x.index(foo) 
        
                number = x[start:end].strip(' ')
            
    except: 
        number ='0' 
    
    print ("clean           "+number)

Output:

raw             1x 80 3x 4x
clean           80
raw             4x 3x 80 1x
clean           80
raw             1x 80 4x 
clean           0
raw             4x 80 1x
clean           0
raw             80 3x 4x
clean           0
raw             4x 3x 80
clean           0
raw             80 4x
clean           0
raw             4x 80
clean           0

Answer 1

this look like a job for regular expressions

>>> import re
>>> text="""1x 80 3x 4x
4x 3x 80 1x
1x 80 4x
4x 80 1x
80 3x 4x
4x 3x 80
80 4x
4x 80""".splitlines()
>>> text
['1x 80 3x 4x', '4x 3x 80 1x', '1x 80 4x', '4x 80 1x', '80 3x 4x', '4x 3x 80', '80 4x', '4x 80']
>>> for t in text:
        res = re.search("((1|4). )?(3. )?(?P<result>[^ ]+)(3. )?((1|4). )?",t)
        print(f"raw: {t!r}\nclean: {res['result']!r}")

    
raw: '1x 80 3x 4x'
clean: '80'
raw: '4x 3x 80 1x'
clean: '80'
raw: '1x 80 4x'
clean: '80'
raw: '4x 80 1x'
clean: '80'
raw: '80 3x 4x'
clean: '80'
raw: '4x 3x 80'
clean: '80'
raw: '80 4x'
clean: '80'
raw: '4x 80'
clean: '80'
>>> 

here . represent any character, (a|b) is a or b for expresions a and b, (...)? means that the inside is optional so ((1|4). )? mean 1 or 4 alongside one character plus one space and is optional, similar for the others, and for (?P<result>[^ ]+) , (?P<name>...) means that is a group named name , [^ ] is any character but space and the plus sign is that we want one or more

UPDATE:

import re

POSRE = "(?P<pos>(?:[1-9](?:[\.\: ] ?)?[a-zA-Z](?: |$)))"

def extrator(rawtext):
    result = filter(None,map(str.strip,re.split(POSRE,rawtext)))
    result = [(x,int(x[0]) if re.match(POSRE,x) else None) for x in result]
    pos=[n for x,n in result if n]
    if sorted(pos)!=pos:
        result = list(reversed(result))
    final = [x for x,p in result if p is None]
    if len(final)==2 and 6 not in pos:
        a,b = final
        final = [a,*b.split()]
    elif len(final)<3:
        final.extend([None]*(3-len(final)))
    return final

So the main thing here is identify those position marker which structure is know, for that I device that regular expression were we check if its a single number ( [1-9] ) optionally follow by a . or : or ( (?:[\.\: ]?)? ) then a letter ( [a-zA-Z] ) and then another space or the end of the string ( (?: |$) ). The (?:...) means that is a not capturing group, for more detail on those check the documentation linked above...

We use that in re.split to split the text into its matching and not matching parts which are then strip out of their surrounding spaces characters and filter out those that turn out to be empty.

We follow that by identifying what is their position if they are a matching string or None if not.

Then is just a couple simple check, like in what order they came and reversed it if needed so we always return in the same order and extract what we need in final , check for the final case and adjust accordingly and done.

and a little test

text="""1. A  $67  4. A  69.000kg  6. A  12sec  8. B  9. B
9B 8B 22 sec 6A 75.000kg 4b  $80 1b
10 Mrd 3: A 4: A  50 .379 6: A   7:19   8: B 9: D
9a 8b 10 6b 60000 4a 3 b 50
10 Mrd 4: A  50 .379 6: A   7:19   8: B 9: D
9a 8b 10 6b 60000 4a  50
9B 8B 22 sec 6A 75.000kg 4b 3b  $80 1b
1. A  $67  3. A    69.000kg  6. A  12sec  8. B  9. B
1. A  $67  3. A  4a  69.000kg   12sec  8. B  9. B
9a 8b 6b 4a 3 b 50 1b""".splitlines()

for t in text:
    print(f"raw: {t!r}\nresult: ",extrator(t) )
    print()

which give us

raw: '1. A  $67  4. A  69.000kg  6. A  12sec  8. B  9. B'
result:  ['$67', '69.000kg', '12sec']

raw: '9B 8B 22 sec 6A 75.000kg 4b  $80 1b'
result:  ['$80', '75.000kg', '22 sec']

raw: '10 Mrd 3: A 4: A  50 .379 6: A   7:19   8: B 9: D'
result:  ['10 Mrd', '50 .379', '7:19']

raw: '9a 8b 10 6b 60000 4a 3 b 50'
result:  ['50', '60000', '10']

raw: '10 Mrd 4: A  50 .379 6: A   7:19   8: B 9: D'
result:  ['10 Mrd', '50 .379', '7:19']

raw: '9a 8b 10 6b 60000 4a  50'
result:  ['50', '60000', '10']

raw: '9B 8B 22 sec 6A 75.000kg 4b 3b  $80 1b'
result:  ['$80', '75.000kg', '22 sec']

raw: '1. A  $67  3. A    69.000kg  6. A  12sec  8. B  9. B'
result:  ['$67', '69.000kg', '12sec']

raw: '1. A  $67  3. A  4a  69.000kg   12sec  8. B  9. B'
result:  ['$67', '69.000kg', '12sec']

raw: '9a 8b 6b 4a 3 b 50 1b'
result:  ['50', None, None]

UPDATE 2

here is a version which identify which one is the data we got given a couple of assumption such as:

• there is only position markers and data, the data is only a positions 2, 5 and 7
• the previous regular expression can identify those position markers
• any of those can go missing
• and there is no space characters in the data so in case any of the relevant positions markers is missing and less data than expect is found then one of those may be group in one of the data points extracted and thus can be safely str.split , if that is not the case, adjust in those parts accordingly.

this result in a rather lengthy case by case check I hope is self explanatory and return a dictionary that said who is who.

surely this can be refined, but no refinement had come to mind.

def extrator(rawtext):
    fil  = filter(None,map(str.strip,re.split(POSRE,rawtext)))
    proc = [(x,int(x[0]) if re.match(POSRE,x) else None) for x in fil] #process raw data
    pos  = [p for x,p in proc if p is not None ] #position markers presents
    if sorted(pos)!=pos:
        proc = list(reversed(proc))        
    data = [x for x,p in proc if p is None]
    pos = {p:i for i,(x,p) in enumerate(proc) if p is not None } #pos marker:index of it
    #print(f"{proc=}")
    if len(data)==3:
        return dict(zip((2,5,7),data))
    #from here, a,b,c will represent data in position 2,5 and 7 respectively
    elif len(data)==2:
        a,b = data
        #c = None
        if 3 in pos or 4 in pos:
            if 6 in pos:
                #one of 2, 5 or 7 is missing
                i = proc.index( (a,None) )
                i34 = pos[3] if 3 in pos else pos[4]
                if i < i34:
                    #a is 2, b is 5 or 7
                    j = proc.index( (b,None) )
                    if j < pos[6]:
                        #7 is missing
                        c = None
                    else:
                        #5 is missing
                        b,c = None,b
                else:
                    #2 is missing, a is 5 thus b is 7
                    a,b,c = None,a,b
            else:
                #a is 2, b may be 5 or 7 or both
                t = b.split()
                if len(t) == 2:
                    #b was both
                    b,c = t
                elif len(t) == 1:
                    #b is 5 or 7
                    print("either 5 or 7 is missing, picked 7 as missing")
                    c = None
                else:
                    #b was split into more than 2 parts
                    raise RuntimeError("unknow case 1")
        else:
            #3 and 4 are missing
            if 6 in pos:
                #a may be 2 or 5 or both, b is 7
                c = b
                t = a.split()
                if len(t) == 2:
                    #a was both
                    a,b = t
                elif len(t) == 1:
                    print("either 2 or 5 is missing, picked 5 as missing")
                    b = None
                else:
                    #a was split into more than 2 parts
                    raise RuntimeError("unknow case 2")
            else:
                raise RuntimeError("Fatal error: 2 data points with no marker in between")
        return dict(zip((2,5,7),(a,b,c)))
    elif len(data)==1:
        a = data[0]
        i = proc.index( (a,None) )
        #b,c = None, None
        if 3 in pos or 4 in pos:
            i34 = pos[3] if 3 in pos else pos[4]
            if 6 in pos:
                #only one of 2,5 or 7 are present
                if i < i34:
                    #a is 2 the rest is missing
                    b,c = None, None
                elif i < pos[6]:
                    #a is 5
                    a,b,c = None, a, None
                else:
                    #a is 7
                    a,b,c = None, None, a 
            else:
                #a is 2 or a is 5 or 7 or both
                if i < i34:
                    #a is 2, the rest is missing
                    b,c = None, None
                else:
                    #2 is missing, a is 5 or 7 or both 
                    a,b = None, a
                    t = b.split()
                    if len(t) == 2:
                        b,c = t
                    elif len(t) == 1:
                        print("either 5 or 7 is missing, picked 7 as missing")
                        c = None
                    else:
                        raise RuntimeError("unknow case 3")
        else:
            #3 and 4 are missing
            if 6 in pos:
                if pos[6] < i:
                    #a is 7, the rest is missing
                    a,b,c = None, None, a
                else:
                    #7 is missing, a is 2 or 5 or both
                    c = None
                    t = a.split()
                    if len(t) == 2:
                        a,b = t
                    elif len(t) == 1:
                        print("either 2 or 5 is missing, picked 5 as missing")
                        b = None
                    else:
                        raise RuntimeError("unknow case 4")
            else:
                #a is 2, 5 or 7 or any combination of them
                t = a.split()
                if len(t) == 3:
                    a,b,c = t
                elif len(t) == 2:
                    print("one of 2, 5 or 7 is missing, picked 7 as missing")
                    a,b = t
                    c = None
                elif len(t) == 1:
                    print("only one of 2, 5 or 7 is present, picked 2 as present")
                    b,c = None, None
                else:
                    raise RuntimeError("unknow case 5")
        return dict(zip((2,5,7),(a,b,c)))
    elif len(data) == 0:
        return dict.fromkeys( (2,5,7) )
    else:
        raise RuntimeError("unknow case 6: more than 3 data points")


def test():
    text="""1. A  $67  4. A  69.000kg  6. A  12sec  8. B  9. B
9B 8B 22 sec 6A 75.000kg 4b  $80 1b
10 Mrd 3: A 4: A  50 .379 6: A   7:19   8: B 9: D
9a 8b 10 6b 60000 4a 3 b 50
10 Mrd 4: A  50 .379 6: A   7:19   8: B 9: D
9a 8b 10 6b 60000 4a  50
9B 8B 22 sec 6A 75.000kg 4b 3b  $80 1b
1. A  $67  3. A    69.000kg  6. A  12sec  8. B  9. B
1. A  $67  3. A  4a  69.000kg   12sec  8. B  9. B
9a 8b 6b 4a 3 b 50 1b
9 a 8b 6 b 55 4a 3 b 1b
9a 8:b 777 6 b 4.a 3 b 1b
9a 8:b 777 6 b 4.a 3 b 55 1b
""".splitlines()

    for t in text:
        print(f"raw: {t!r}\nresult: ",extrator(t) )
        print()

output

>>> test()
raw: '1. A  $67  4. A  69.000kg  6. A  12sec  8. B  9. B'
result:  {2: '$67', 5: '69.000kg', 7: '12sec'}

raw: '9B 8B 22 sec 6A 75.000kg 4b  $80 1b'
result:  {2: '$80', 5: '75.000kg', 7: '22 sec'}

raw: '10 Mrd 3: A 4: A  50 .379 6: A   7:19   8: B 9: D'
result:  {2: '10 Mrd', 5: '50 .379', 7: '7:19'}

raw: '9a 8b 10 6b 60000 4a 3 b 50'
result:  {2: '50', 5: '60000', 7: '10'}

raw: '10 Mrd 4: A  50 .379 6: A   7:19   8: B 9: D'
result:  {2: '10 Mrd', 5: '50 .379', 7: '7:19'}

raw: '9a 8b 10 6b 60000 4a  50'
result:  {2: '50', 5: '60000', 7: '10'}

raw: '9B 8B 22 sec 6A 75.000kg 4b 3b  $80 1b'
result:  {2: '$80', 5: '75.000kg', 7: '22 sec'}

raw: '1. A  $67  3. A    69.000kg  6. A  12sec  8. B  9. B'
result:  {2: '$67', 5: '69.000kg', 7: '12sec'}

raw: '1. A  $67  3. A  4a  69.000kg   12sec  8. B  9. B'
result:  {2: '$67', 5: '69.000kg', 7: '12sec'}

raw: '9a 8b 6b 4a 3 b 50 1b'
result:  {2: '50', 5: None, 7: None}

raw: '9 a 8b 6 b 55 4a 3 b 1b'
result:  {2: None, 5: '55', 7: None}

raw: '9a 8:b 777 6 b 4.a 3 b 1b'
result:  {2: None, 5: None, 7: '777'}

raw: '9a 8:b 777 6 b 4.a 3 b 55 1b'
result:  {2: '55', 5: None, 7: '777'}

>>> 

Answer 2

If I understand you goals correctly, it seems to me like you are far over-complicating the process. I wrote a function which splits the input string to a list and checks whether each segment meets the formatting of 1x , 2x , or 3x . Check it out and let me know if it's not what you need.

# we use regex to check for matches with the format
import re

#list of strings
x_list = ["1x 80 3x 4x","4x 3x 80 1x"]
for x in x_list:
    print(find_substr(x))

def find_substr(x):
    # break on spaces into a list
    seg = x.split(" ")
    # check each word for the desired format
    for i in range(len(seg)):
        for j in [1,3,4]:
            if re.search(str(j)+".", seg[i]) is None:
                # this word does not fit the format, so it's the substring
                return seg

Answer 3

Changed the code up a little to be more readable. I didn't know if this is what you wanted but it works. Let me know if you have any questions. I would be happy to help

for x in x_list:
    print ("raw             "+x)

    try:
        # splits the string into a list, separating on spaces (e.g ['1x', '80', '3x', '4x'])
        y = x.split(" ")

        # a is the substring that you are checking in the list
        a = '80'
        if a in y:
            index = y.index(a)
            number = y[index]

    except: 
        number ='0' 
    
    print ("clean           "+number)