How to identify invalid pattern using regx? - regex

Question

I have a dataset such as below:
import pandas as pd
dic={"ID":[1,2,3,4,5,6],
"Size":["3-4mm","12mm",math.nan,"1 mm","1mm, 2mm, 3mm","13*18mm"]}
dt = pd.DataFrame(dic)
so, the dataset is:
ID Size
1 3-4mm
2 12mm
3 NaN
4 1 mm
5 1mm, 2mm, 3mm
6 13*18mm
In the column Size, i should have only 3 valid patterns, and anything except these 3 are invalid. These 3 pattern are as below
3-4mm (int-intmm)
NaN
4mm (intmm)
I am wondering how can i have function which specifies the ID of the rows which has invalid size pattern?
So, in my example:
ID
4
5
6
The reason is their size is not in valid format.
I have no preference for the solution, but i guess the most easiest solution comes from Regx

Answer 1

using #CodeManiac's pattern, you can pass it to series.str.contains() and pass the na parameter as True since it is a actual NaN:
dt.loc[~dt.Size.str.contains('^(?:\d+-\d+mm|\d+mm)$',na=True),'ID']
3 4
4 5
5 6
Details:
executing: dt.Size.str.contains('^(?:\d+-\d+mm|\d+mm)$')
0 True
1 True
2 NaN
3 False
4 False
5 False
pass na=True to fill NaN as True:
dt.Size.str.contains('^(?:\d+-\d+mm|\d+mm)$',na=True)
0 True
1 True
2 True
3 False
4 False
5 False
Then use invert ~ to invert True as False and vice versa since we want False values and call the ID column under df.loc[]

Answer 2

The function that returns 'ID'-s of rows with invalid value in 'Size' column:
import re # standard Python regular expressions module
def get_invalid(dt):
return dt[dt['Size'].apply(lambda r: re.match(r'^\d+-\d+mm|nan|\d+mm$', str(r), re.MULTILINE) is None)]['ID']
Output:
3 4
4 5
5 6
Name: ID, dtype: int64