Generate histogram from regex matches - sql

Question

Sorry if this is an obvious question. I'm quite new to SQL and couldn't manage to adapt other examples out there to my needs.
I have a table (Postgres 9.3) defined as:
CREATE TABLE scripts (
id SERIAL PRIMARY KEY,
name VARCHAR(256) NOT NULL,
content TEXT NOT NULL);
The content column contains the content of various scripts. I'm interested in counting how many times distinct function calls occur in these scripts.
I've managed to construct a query that runs a regex over the contents, and pulls out all the function calls (as funcs)
SELECT id, name, regexp_matches(LOWER(content), '(\w+\.\w+)\(', 'g') AS funcs
FROM scripts
GROUP BY id, name, funcs;
The output looks something like
1, myscript, {class.m1}<br>
2, otherscript, {class_b.method4}<br>
2, otherscript, {class.m1}<br>
3, last_script, {classname.method2}<br>
3, last_script, {class.m1}<br>
3, last_script, {class_b.method4}<br>
I would really like to turn this into a table that shows a tally of each distinct function. Something like
class.m1, 3
class_b.method4, 2
classname.method2, 1
This is what I have so far:
SELECT COUNT(DISTINCT funcs) FROM (
SELECT tsr_id, name, regexp_matches(LOWER(content), '(\w+\.\w+)\(', 'g') AS funcs
FROM tsr_conf.rules
GROUP BY tsr_id, name, funcs
) x
But unfortunately it just gives me the total count of distinct functions. Any advice on how to count the occurances of each distinct function would be most appreciated!

Answer 1

Given what your first query returns, a group by should do what you want:
SELECT funcs, COUNT(*)
FROM (SELECT tsr_id, name, regexp_matches(LOWER(content), '(\w+\.\w+)\(', 'g') AS funcs
FROM tsr_conf.rules
GROUP BY tsr_id, name, funcs
) x
GROUP BY funcs;
You could actually write this more simply as:
SELECT regexp_matches(LOWER(content), '(\w+\.\w+)\(', 'g') AS funcs, COUNT(DISTINCT tsr_id, name)
FROM tsr_conf.rules
GROUP BY funcs;