The top voted answer uses COLLATE UTF8_GENERAL_CI.
There is also an answer with lower, some comment mentions lower, it has a few comments related to that.
@Col. -- Admittedly, this is less than ideal for indexed columns, but it will work for a structure which is already in place. I've also found that case-insensitive searches are more often on columns which are not indexed anyway. — cwallenpooleMay 20 '10 at 18:48
(Body LIKE '%[^a-zA-Z]##text?matrice##[^a-zA-Z]%' COLLATE Latin1_General_CI_AI) OR (Title LIKE '%[^a-zA-Z]##text##[^a-zA-Z]%' COLLATE Latin1_General_CI_AI)
@Martin yeah, I think mathematics has the benefit of having way less rows then SO main. Combined with the awesome hardware SQL Server runs on a lot of things that might go south on SO does work for all other sites.
I am trying to find out the tags that appear frequently together for a given tag.
for e.g pandas will appear frequently with python3 and nltk will appear together with nlp and so on.
So by running a query for pandas, I want to get the more frequently used tags with pandas.