This Postgres extension introduces two new data types inthstore and floathstore wich depends on postgres
hstore extension.
These types are used to hold integer and float values respectively.
numhstore comes with certain aggregate and helper functions.
##Operators
SELECT 'a=>3, b=>2'::inthstore + 'a=>2, c=>5'::inthstore;
inthstore
------------------------------
"a"=>"5", "b"=>"2", "c"=>"5"
SELECT 'a=>3, b=>2'::inthstore - 'a=>2, c=>5'::inthstore;
inthstore
-------------------------------
"a"=>"1", "b"=>"2", "c"=>"-5"
SELECT 'a=>3, b=>2'::inthstore * 'a=>2, c=>5'::inthstore;
inthstore
--------------------------------
"a"=>"6", "b"=>NULL, "c"=>NULL
SELECT 'a=>3, b=>2'::inthstore / 'a=>2, c=>5'::inthstore;
floathstore
----------------------------------------------------------
"a"=>"1.5000000000000000", "c"=>"0.00000000000000000000"
SELECT 'a=>3, b=>2'::inthstore + 3;
inthstore
--------------------
"a"=>"6", "b"=>"5"
SELECT 'a=>3, b=>2'::inthstore - 3;
inthstore
---------------------
"a"=>"0", "b"=>"-1"
SELECT 'a=>3, b=>2'::inthstore * 3;
inthstore
--------------------
"a"=>"9", "b"=>"6"
SELECT 'a=>3, b=>2'::inthstore / 3;
floathstore
--------------------------------------------------------------
"a"=>"1.00000000000000000000", "b"=>"0.66666666666666666667"
##Functions
SELECT array_count(ARRAY['a','a','b']);
inthstore
----------------------
"a"=>"2", "b"=>"1"
SELECT hstore_sum_up('foo=>10, bar => 15'::inthstore);
bigint
---------
25
SELECT hstore_sum_up('foo=>2.5, bar => 4.5'::floathstore);
numeric
----------
7.0
SELECT array_add(Array['foo','foo','bar'], Array[2,3,5]);
inthstore
------------------------
"bar"=>"5", "foo"=>"5"
###SUM
SELECT SUM(store) FROM (VALUES('a=>2, b=>4'),('a=>4, b=>6, c=>6'::inthstore) )t(store);
inthstore
-------------------------------
"a"=>"6", "b"=>"10", "c"=>"6"
###AVG
SELECT AVG(store) FROM (VALUES('a=>2, b=>4'),('a=>4, b=>6, c=>6'::inthstore) )t(store);
floathstore
---------------------------------------------------------------------------------
"a"=>"3.0000000000000000", "b"=>"5.0000000000000000", "c"=>"6.0000000000000000"
(1 row)