I'm trying to use the icu RuleBasedCollator in python. In my code I specify a rule wherby "ä" should sort before "a" as a secondary (accent) difference

from icu import RuleBasedCollator

l=["a","ä"]
rbc = RuleBasedCollator('\n&ä<<a')
sorted(l, key=rbc.getSortKey)

However, the output of the sorted is:

['a', 'ä']

I expected: ['ä','a'] What did I do wrong?

Many thanks

I'm trying to use the icu RuleBasedCollator in python. In my code I specify a rule wherby "ä" should sort before "a" as a secondary (accent) difference

from icu import RuleBasedCollator

l=["a","ä"]
rbc = RuleBasedCollator('\n&ä<<a')
sorted(l, key=rbc.getSortKey)

However, the output of the sorted is:

['a', 'ä']

I expected: ['ä','a'] What did I do wrong?

Many thanks

• python
• collation
• icu

1 Answer 1

It appears that the difference between a and ä is considered primary. Using [before 1] you can achieve the expected result.

from icu import RuleBasedCollator

l=["a","ä"]
rbc = RuleBasedCollator('&[before 1]a < ä')
print(sorted(l, key=rbc.getSortKey))

To read more