Question

lets say we have a table BNK_ACCT_GRP_ST with these columns:

Column Name                          ID   Pk    Null? Data Type
BNK_ACCT_GRP_ST_ID1                  1          N     NUMBER (15)
BNK_ACCT_ID                          2          N     NUMBER (15)
BNK_ACCT_GRP_ID                      3          N     NUMBER (15)
BNK_ACCT_GRP_ST_CD                   4          N     NUMBER (4)
BNK_ACGRP_ST_SRC_EVNT_TYP_CD         8          N     NUMBER (4)
BNK_ACGRP_ST_SRC_TRAN_ID             9          N     NUMBER (15)
FNCL_GRP_ID                          10         N     NUMBER (15)
CREN_DT                              12         N     DATE
LAST_UPD_DT                          13         N     DATE
PCSG_DT                              14         N     DATE
BNK_ACCT_GRP_ST_SEQ_NO               15         N     NUMBER (15)

its populated with bnk_accts data :

example#1:

BNK_ACCT_GRP_ST_ID  BNK_ACCT_ID BNK_ACCT_GRP_ID BNK_ACCT_GRP_ST_CD BNK_ACGRP_ST_SRC_EVNT_TYP_CD FNCL_GRP_ID BNK_ACCT_GRP_ST_SEQ_NO
2282                150627009   1724            4                  150                          111111111   2
2283                150627009   1440            2                  149                          111111112   1
1908                150627009   1725            2                  134                          111111111   1
1906                150627009   1441            2                  135                          111111111   7

and example #2:

BNK_ACCT_GRP_ST_ID  BNK_ACCT_ID BNK_ACCT_GRP_ID BNK_ACCT_GRP_ST_CD BNK_ACGRP_ST_SRC_EVNT_TYP_CD FNCL_GRP_ID BNK_ACCT_GRP_ST_SEQ_NO
2364                150275031   1435            2                  114                          133333333   3
2365                150275031   1436            7                  116                          133333333   2
1902                150275031   1435            1                  72                           133333333   2
1903                150275031   1435            2                  36                           133333333   1
1904                150275031   1436            5                  74                           133333333   1

where:

• BNK_ACCT_GRP_ST_ID is autoincrement field,
• BNK_ACCT_ID is actualy bank account # id,
• BNK_ACCT_GRP_ID is grp id (which always increasing),
• BNK_ACCT_GRP_ST_CD, BNK_ACGRP_ST_SRC_EVNT_TYP_CD and BNK_ACGRP_ST_SRC_TRAN_ID are some id’s for trans purposes,
• FNCL_GRP_ID – id of the bankiing insitution where this bank acc belongs to

what i need : to create a query to pull latest BNK_ACCT_GRP_ID for each BNK_ACCT_ID based on FNCL_GRP_ID. here one bank acc can belongs to one or more then one fncl instituion, so the query should return this output:

for example#1:

FNCL_GRP_ID  BNK_ACCT_ID  BNK_ACCT_GRP_ID
111111111    150627009    1725
111111112    150627009    1440

for example#2:

FNCL_GRP_ID  BNK_ACCT_ID  BNK_ACCT_GRP_ID
133333333    150275031    1436

I have tried a lot of the things lately, such as a combo of

(select distinct BNK_ACCT_ID,FNCL_GRP_ID) join to rank() over (partiton  by ) 

but with no luck.

Answer 1

Try this:

WITH BankData AS
(
    SELECT  a.*, 
            ROW_NUMBER()
              OVER(PARTITION BY fncl_grp_id, bnk_acct_id ORDER BY bnk_acct_grp_id DESC) AS Position
      FROM bnk_acct_grp_st a
)
SELECT  *
  FROM  BankData
 WHERE  Position = 1