Hive中的regexp_extract参数

Question

What does the argument in the curly brackets do in the following segment of code?

regexp_extract(col_value, '^(?:([^,]*)\,?){1}', 1) Id,  
regexp_extract(col_value, '^(?:([^,]*)\,?){2}', 1) Score,  
regexp_extract(col_value, '^(?:([^,]*)\,?){9}', 1) DisplayName,

Answer 1

As you can read here , curly brackets contain how many times the preceding token, in this case a non-capturing group , may repeat.

The group contains a (possibly empty) capturing group made of non-comma characters, followed by an optional comma. Since there is only one number in curly brackets, the non-capturing group must repeat exactly that number of times.

I don't know why the comma should be escaped by a backslash. It seems to me that the backslash is not necessary.

Caveat: I don't know Hadoop or Hive, all my knowledge of regexp_extract comes from this page .

The purpose of these regexes is to match the first, second and ninth element in a comma-separated list, where the capturing group #1 (selected by the third argument of regexp_extract ) returns only its last occurrence. Of course the comma is not really optional, except after the last element.