Question

If I have this in a Postgresql 9.1 column:

foo foo <th id="ddd"> foo foo <th id="www"> foo

And I want it to update to this:

foo foo <th> foo foo <th> foo

I’ve tried regex_replace, but I have not succeeded.

Answer 1

Assuming you have a table like this:

CREATE TABLE table1
(
  a character varying NOT NULL,
  ...
)

You can use the following regexp_replace:

update table1 set a = regexp_replace(a, '(.*?)<(\w+)\s+.*?>(.*)', '\1<\2>\3', 'g');

The 'g' flag indicates to replace all matching patterns, not only the first one.

With this input:

foo foo <th id="ddd"> foo foo <th id="www"> foo<div id="hey">

I get the following ouput:

foo foo <th> foo foo <th> foo<div>