Equivalent DML Program, involves Max with Rank 1. Perf Issues.

Question

Hoping that someone could shed some light on how one might be able to perform the same using OLAP DML instead of SQL with decent performance.
PRICE, DT and TM are dimensions. EQUITY_VP_VIEW is a cube with a single measure based on SUM, volume. I am able to obtain the price which contains the maximum volume on a given day. The SQL Query is straight forward. As shown below and returns instantly.
SELECT *
FROM
(SELECT d.end_date date_id,
p.short_description price,
e.volume,
row_number() over (partition BY d.end_date order by e.volume DESC) rank
FROM EQUITY_VP_VIEW e,
DT_H_VIEW d,
ES_PRICE_H_VIEW p,
TM_H_VIEW t
WHERE e.dt = d.dim_key
AND e.es_price = p.dim_key
AND d.level_name = 'DAY'
AND e.tm = t.dim_key
AND t.level_name = 'ALL_TIME'
AND p.level_name = 'PRICE'
ORDER BY d.end_date
WHERE rank = 1
Trying to perform the same via DML. See below. This takes 84 seconds to return. Would appreciate any input.
argument _tm TM
argument _dt DT
variable _lg number
variable _ret number
_ret = na
limit es_price_new to ES_PRICE_NEW_LEVELREL eq 'PRICE'
lg = largest(EQUITYVP_VOLUME(TM tm DT dt))
if _lg ne na
then do
limit ES_PRICE_NEW keep EQUITY_VP_VOLUME(TM tm DT dt) eq _lg ifnone SKIP1
ret = convert(ESPRICE_NEW_SORT_ATTR, number)
doend
SKIP1:
return _ret                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                                   

Answer 1

Have you tried defining a calculated measure that returns the rank? You should be able to do it in AWM and the measure expression would look something like this:
RANK() OVER HIERARCHY (
  ES_PRICE_NEW.HIER_NAME
  ORDER BY EQUITY_VP.VOLUME ASC NULLS LAST
  WITHIN LEVEL)You can then use the generated formula in OLAP DML or the column in SQL as you need. If you want to use raw OLAP DML, then consider using the "LIMIT TO TOP n BASEDON expression" command instead of a combination of LARGEST and LIMIT. Also, drop the "EQ" in your first limit before the 'PRICE'.
limit es_price_new to ES_PRICE_NEW_LEVELREL 'PRICE'
limit es_price_new keep top 1 basedon EQUITY_VP_VOLUME(TM _tm DT _dt)