Regular expression

mreazo · 10-07-2021, 07:48 PM

I'm not quite sure if this is the correct place to ask this question, but I'm working on a script that will identify the case of a string so upper case, lower case, and mixed case. I have upper and lower figured out but I'm ramming my head against the wall on mixed case.

If I input just lower case like "table" it will identify as lower and mixed

If I input just upper case like "TABLE" it will identify as upper and mixed

If I input mixed case like "Table" it will identify as mixed.

I guess my question is what would be a correct regular expression that only identifies a mixed case string like "TaBle" and not "table" and "TABLE"

Right now I'm using-

Uppercase- ^[[:upper:]]+$

Lowercase- ^[[:lower:]]+$

Mixed- ^[[:alpha:]]+$

Sorry for the long post, I'd appreciate any help!!

sundialsvcs · 10-07-2021, 10:34 PM

Please: "show us the actual code."

In whatever scripting language it may be.

scasey · 10-07-2021, 10:37 PM

:alpha: is a superset containing both :upper: and :lower:, so the responses you’re getting are correct, technically.
Your logic needs to not test for mixed if either upper or lower are true.
It’s not a regular expression issue.

Edit: yes. We need to see the logic to be more specific

syg00 · 10-07-2021, 11:04 PM

For mixed the test must be not Uppercase AND not Lowercase.

ondoho · 10-07-2021, 11:26 PM

Just to clarify what others already said:

Quote:

Originally Posted by mreazo

Mixed- ^[[:alpha:]]+$

This is actually wrong. [:alpha:] means any sort of letter, it does not identify mixed case.
As the last 2 posts said, the logic you need is sth like this:
test if string is [:alpha:] and NOT [:upper:] and NOT [:lower:]

pan64 · 10-08-2021, 01:47 AM

You have to write a regexp which contains both upper and lowercase letters, but actually that is order dependent, so it will look like:

Code:

([[:upper:]][[:lower:]])|([[:lower:]][[:upper:]])
|           < is OR
()          < grouping

allend · 10-08-2021, 02:07 AM

If the script is for the bash shell, then there is a useful parameter expansion

Code:

#!/bin/bash

v1="TABLE"
v2="table"
v3="Table"

for v in $v1 $v2 $v3; do
  if [[ $v == ${v@U} ]]; then
    echo "$v is Uppercase"
  elif [[ $v == ${v@L} ]]; then
    echo "$v is Lowercase"
  else
    echo "$v is Mixedcase"
  fi
done

boughtonp · 10-08-2021, 07:52 AM

Quote:

Originally Posted by mreazo

Sorry for the long post, I'd appreciate any help!!

Your post isn't long, but as mentioned you should include the actual function you currently have (rather than describing it), as well as clearly identifying what language[s] are involved - regular expressions syntax and functionality varies.

There was mention of "not upper and not lower" as a valid test for mixed case - in response I will point out that you will probably want to test how such a function responds if the input is an empty string.

Other inputs worth considering include:

Code:

A_
a_
12345
!"£$%
café straße

Good programming involves considering all possible inputs and behaving appropriately.

MadeInGermany · 10-08-2021, 02:17 PM

"All uppercase" and "All lowercase" are border cases of "All alpha" - need to be sorted out.
The following is like post #7 but spots the "Not alpha" case (including the "empty" case):

Code:

v1="TABLE"
v2="table"
v3="Table"
v4="A4"
v5=""

for v in "$v1" "$v2" "$v3" "$v4" "$v5"
do
  if [[ $v =~ ^[[:upper:]]+$ ]]
  then
    echo "'$v' - Uppercase"
  elif [[ $v =~ ^[[:lower:]]+$ ]]
  then
    echo "'$v' - Lowercase"
  elif [[ $v =~ ^[[:alpha:]]+$ ]]
  then
    echo "'$v' - Mixed"
  else
    echo "'$v' - Not alpha"
  fi
done