Applying regex to pandas column based on different pos of same character

user_10829044 0
Viewed 581

I have a dataframe like as shown below

tdf = pd.DataFrame({'text_1':['value: 1.25MG - OM - PO/TUBE - ashaf', 'value:2.5 MG - OM - PO/TUBE -test','value: 18 UNITS(S)','value: 850 MG - TDS AFTER FOOD - SC (SUBCUTANEOUS) -had', 'value: 75 MG - OM - PO/TUBE']})

I would like to apply regex and create two columns based on rules given below

col val should store all text after value: and before first hyphen

col Adm should store all text after third hyphen

I tried the below but it doesn't work accurately

tdf['text_1'].str.findall('[.0-9]+\s*[mgMG/lLcCUNIT]+')

enter image description here

pandas dataframe python regex string
3 Answers

Series.str.extract

tdf['text_1'].str.extract(r'^value:\s?([^-]+)(?:\s-.*?-\s)?([^-]*)(?:\s|$)')

             0                  1
0       1.25MG            PO/TUBE
1       2.5 MG            PO/TUBE
2  18 UNITS(S)                   
3       850 MG  SC (SUBCUTANEOUS)
4        75 MG            PO/TUBE

Regex details:

  • ^ : Assert position at start of line
  • value: : Matches character sequence value:
  • \s?: Matches any whitespace character between zero and one time
  • ([^-]+) : First capturing group matches any character except - one or more times
  • (?:\s-.*?-\s)? : Non capturing group match between zero and one time
    • \s: Matches single whitespace character
    • - : Matches character -
    • .*? : Matches any character between zero and unlimited times but as few times as possible
    • - : Matches character -
    • \s : Matches single whitespace character
  • ([^-]*) : Second capturing group matches any character except - zero or more times
  • (?:\s|$) : Non capturing group
    • \s- : Matches single whitespace character
    • | : Or switch
    • $ : Assert position at the end of line

See the online Regex demo

user_128331660

With your shown samples, could you please try following.

tdf[["val", "Adm"]] = tdf["text_1"].str.extract(r'^value:\s?(\S+(?:\s[^-]+)?)(?:\s-\s.*?-([^-]*)(?:-.*)?)?$', expand=True)
tdf

Online demo for above regex

Output will be as follows.

                                                    text_1          val                  Adm
0                     value: 1.25MG - OM - PO/TUBE - ashaf       1.25MG             PO/TUBE 
1                        value:2.5 MG - OM - PO/TUBE -test       2.5 MG             PO/TUBE 
2                                       value: 18 UNITS(S)  18 UNITS(S)                  NaN
3  value: 850 MG - TDS AFTER FOOD - SC (SUBCUTANEOUS) -had       850 MG   SC (SUBCUTANEOUS) 
4                              value: 75 MG - OM - PO/TUBE        75 MG              PO/TUBE

Explanation: Adding detailed explanation for above.

^value:\s?       ##Checking if value starts from value: space is optional here.
(\S+             ##Starting 1st capturing group from here and matching all non space here.
  (?:\s[^-]+)?   ##In a non-capturing group matching space till - comes keeping it optional.
)                ##Closing 1st capturing group here.
(?:\s-\s.*?-     ##In a non-capturing group matching space-space till - first occurrence.
  ([^-]*)        ##Creating 2nd capturing group which has values till next - here.
  (?:-.*)?       ##In a non capturing group from - till end of value keeping it optional.
)?$              ##Closing non-capturing group at the end of the value here.
user_58665800

You can use

tdf[["val", "Adm"]] = tdf["text_1"].str.extract(r'^val:\s*([^-]*?)(?:\s*-[^-]*-\s*(.*))?$', expand=True)
# => >>> tdf
                                             text_1          val  \
0                        val: 1.25MG - OM - PO/TUBE       1.25MG   
1                         val:2.5 MG - OM - PO/TUBE       2.5 MG   
2                                  val: 18 UNITS(S)  18 UNITS(S)   
3  val: 850 MG - TDS AFTER FOOD - SC (SUBCUTANEOUS)       850 MG   
4                         val: 75 MG - OM - PO/TUBE        75 MG   


0            PO/TUBE  
1            PO/TUBE  
2                NaN  
3  SC (SUBCUTANEOUS)  
4            PO/TUBE

See the regex demo.

Details:

  • ^val: - val: at the start of string (if val: is not always at the start of the string, remove ^ anchor)
  • \s* - zero or more whitespaces
  • ([^-]*?) - Group 1: any chars other than - as few as possible
  • (?:\s*-[^-]*-\s*(.*))? - an optional sequence of
    • \s* - zero or more whitespaces
    • -[^-]*- - a -, any zero or more chars other than -, and then a -
    • \s* - zero or more whitespaces
    • (.*) - Group 2: the rest of the line
  • $ - end of string.
user_38329700
Related