Hi We put this on the feature list request. There is no indication when something along this line will become available. You might consider implementation of a UDF to improve performance. regards, Martin On 9/27/13 9:18 PM, Budulinku Dejmihrasku wrote:

Hi, i am looking for any idea/suggestion how to do something like Oracle has, i.e.

REGEXP_SUBSTR(source, pattern[, position [, occurrence[, match_parameter]]])

which returns the desired substring.

why i need this: i have vast amount of data to parse - i.e. i need to get all of the many possible [parameter=value] pairs from an url access log in order to feed the dimension tables w/parameter-value from the URL by parsing the string directly in MonetDB

when i write a SQL function to do this (or any string parsing as a matter of fact), it is rather slow when parsing millions of records.

i know that there is the PCRE.MAL and i already use with some success functions INDEX, MATCH or REPLACE but i was not able to find anything similar to the REGEXP_SUBSTR functionality to get the desired substring in one go would most likely help me to get better performance any suggestion? please?

thanks milan

_______________________________________________ users-list mailing list users-list@monetdb.org https://www.monetdb.org/mailman/listinfo/users-list