Extract patterned text from a string with Hive

Question

I have data in a column that looks like this:

• Countryside_Video_-_A18-49_Pub_- Q3 -_Flight_7_18_49_BOTH

• Countryside Video - M18-25 Validated -Q4 - Flight 1

• PremiumBrand_2019_Upfront_Video_-_W18-49_Validated_-_Q4_Flight_1_18_49_FEMALE

• Travel Around the World - W25-54 Validated - Q3 25-54_FEMALE

I need to extract the age and gender value from each string:

• A18-49
• M18-25
• W18-49
• W25-54

It's tricky, because there could be any number of combinations between the letters A,M,F and a number range. The letters signify Age, Male, or Female. The number range is the age range.

From some googling, it looks like I might be able to use a regexp_extract function, but I'm a novice to Hive. Any help on this would be greatly appreciated!

Answer 1

我手头没有 Hive 可以测试，但这可能有效：

select regexp_extract(col, '([AMW][0-9]{2}[-][0-9]{2})', 1)