Converting Daily Snapshots to Ranges in PostgreSQL - postgresql

Question

I have a very large table with years' worth of daily snapshots, showing what the data looks like each day. For the sake of illustration the table looks something like this:
Part Qty Snapshot
---- ---- --------
A 5 1/1/2015
B 10 1/1/2015
A 5 1/2/2015
B 10 1/2/2015
A 6 1/3/2015
B 10 1/3/2015
A 5 1/4/2015
B 10 1/4/2015
I would like to implement a slowly changing data methodology and collapse this data into a form that would look like this (assume current date is 1/4/15)
Part Qty From Thru Active
---- ---- -------- -------- ------
A 5 1/1/2015 1/2/2015 I
B 10 1/1/2015 1/4/2015 A
A 6 1/3/2015 1/3/2015 I
A 5 1/4/2015 1/4/2015 A
I have a function that runs daily so when I capture the latest snapshot, I convert it to this methodology. This function runs once the data is actually loaded into the table with an active flag of 'C' (current), from the giant table (which is actually in DB2).
This works for me moving forward (once I have all past dates loaded), but I'd like to have a way to do this in one fell swoop, for all existing dates and convert the individual snapshot dates into ranges.
For what it's worth, my current method is to run this function for every possible date value. While it's working, it's quite slow, and I have several years worth of history to process as I loop one day at a time.
Tables:
create table main.history (
part varchar(25) not null,
qty integer not null,
from_date date not null,
thru_date date not null,
active_flag char(1)
);
create table stage.history as select * from main.history where false;
create table partitioned.history_active (
constraint history_active_ck1 check (active_flag in ('A', 'C'))
) inherits (main.history);
create table partitioned.history_inactive (
constraint history_active_ck1 check (active_flag = 'I')
) inherits (main.history);
Function to process a day's worth of new data:
CREATE OR REPLACE FUNCTION main.capture_history(new_date date)
RETURNS null AS
$BODY$
DECLARE
rowcount integer := 0;
BEGIN
-- partitioned.history_active already has a current snapshot for new_date
truncate table stage.history;
insert into stage.history
select
part, qty,
min (from_date), max (thru_date),
case when max (thru_date) = new_date then 'A' else 'I' end
FROM
partitioned.history_active
group by
part_qty;
truncate table partitioned.history_active;
insert into partitioned.history_active
select * from stage.history
where active_flag = 'A';
insert into partitioned.history_inactive
select * from stage.history
where active_flag = 'I';
END;
$BODY$
LANGUAGE plpgsql VOLATILE
COST 100;