如何捕获正则表达式的首次出现并使用PostgreSQL插入到字符串中

Question

I'm trying to concatenate the digits from a string that starts with 'CityName' into a separate string. I have the concatenation part. My issue is being able to access the matches from the regex I have a regex in rails that looks like /CityName\\s*(\\d+)/i . I'm super new to regex and it's hard for me to wrap my head around the docs. But I'm assuming that this regex will find any digits after the CityName case intensively. And then it's interpolated if it matches an attribute on my model.

regex = /CityName\s*(\d+)/i
if line_1 =~ regex
  "C#{$1}"
  ...
end

But further along in the execution, it's slowing down because I have to iterate over a lot of records. I have a query in psql that will do that calculations that I need, however I'm having a hard time implementing this regex replacement. My attempts so far look like:

CASE
    when addr.line_1 ~* 'CityName\s*(\d+)' then 'C' || regex_matches('CityName\s*(\d+)')[0]
...

I'm having a hard time finding a solution to grab the first occurrence of the regex match. Thanks for any tips :D

EDIT: I am trying to grab the digits after 'CityName' from a string if that string contains 'CityName' Ultimately I need assistance with the regex and how to contactenate the digits with 'C'

Answer 1

Your question is a bit unclear. Are you trying to add the digits to your selection or to filter records based on them?

If you just want to select them:

Address.select(%q{(regexp_matches(addr.line_1, 'CityName\s*(\d+)'))[1] as digits})
  .map(&:digits)

If you want to filter based on then:

Address.where(%q{addr.line_1 ~ 'CityName\s*(\d+)'}).map &:email
  .map(&:line_1)

Also a few notes:

• Selecting digits case intensively does not really make sense. Digits does not have case.
• PostgreSQL arrays start from 1 instead of 0.

Answer 2

Got it! Was able to finally start to figure out the regex. WHEN addr.line_1 ~* '(?i)CityName\\s*(\\d+)' THEN 'C' || (SELECT (regexp_matches(addr.line_1, '(?i)CityName\\s*(\\d+)'))[1]) WHEN addr.line_1 ~* '(?i)CityName\\s*(\\d+)' THEN 'C' || (SELECT (regexp_matches(addr.line_1, '(?i)CityName\\s*(\\d+)'))[1]) The (?i) allowed for case insensitive matching for CityName and then the concatenation worked. Thank you @ti6on for pointing out the index difference with postgres :D

Answer 3

It seems you need a subquery or a WITH query:

SELECT tbl1.col1, sum(...), min(...) FROM (SELECT ..., CASE ...yourregex stuff... END col1 FROM ...) tbl1 GROUP BY 1;

WITH tbl1 AS (SELECT ..., CASE ...yourregex stuff... END col1 FROM ...) SELECT t.col1, sum(...) FROM tbl1 t GROUP BY 1;

If you need them regulary, you can also create views from the query or create a temp table, then you can use it in queries later.