How to use \b word boundary in pandas str.contains?

Question

Is there an equivalent when using str.contains?

the following code is mistakenly listing "Said Business School" in the category because of 'Sa.' If I could create a wordboundary it would solve the problem. Putting a space after messes this up. I am using pandas, which are the dfs. I know I can use regex, but just curious if i can use strings to make it faster

gprivate_n = ('Co|Inc|Llc|Group|Ltd|Corp|Plc|Sa |Insurance|Ag|As|Media|&|Corporation')
df.loc[df[df.Name.str.contains('{0}'.format(gprivate_n))].index, "Private"] = 1 

Jackelinejackelyn

Sorry, i am using pandas! – Jackelinejackelyn

Just use the regular expression word boundary… – Rittenhouse

@poke: need to use r'\b...' (rawstring). Same old issue that arises with regexes. – Lallygag

Answer 1

A word boundary is not a character, so you can't find it with .contains. You need to either use regex or split the strings into words and then check for membership of each of those words in the set you currently have defined in gprivate_n.

Hipparchus

Word boundaries can be caught with str.contains, when using \\b instead of \b, and / or raw strings. See link and link. – Ballon

@PawelKranzberg: that's actually the same old issue about escaping or raw-string, so use r\b... – Lallygag

This is factually incorrect: \b can be used with str.contains, you just need raw-string: r'\b...' – Lallygag

Answer 2

This is just the same old Python issue in regexes where '\b' should be passed either as raw-string r'\b...'. Or less desirably, double-escaping ('\\b').

So your regex should be:

gprivate_n = (r'\b(Co|Inc|Llc|Group|Ltd|Corp|Plc|Sa |Insurance|Ag|As|Media|&|Corporation)')

Lallygag