Group by Month Using a Column That has Daily Dates - sql

Question

I need to add in an extra 'group by' field to the below code to group by Month in relation to the trade_date field. The code currently results in a table with column headers matching the 4 fields in the beginning 'select' statement. I need to add a 5th that uses the trade_date field to group by month, not day. Any thoughts?
select
TRADEABLE_INSTR_NAME,
TRADING_USERID,
UNDERLYING_SYM_BLOOMBERG,
SUM(DMA_FEE) AS DMA_FEE_SUBTOTAL
FROM
(
SELECT
a.transaction_id,
b.amount,
b.charge_id,
c.charge_type_id,
d.charge_group,
e.account_name,
f.sig_entity_label,
e.trading_userid,
e.exchange_id,
e.tradeable_instr_name,
g.instr_subtype_label,
g.underlying_sym_bloomberg,
g.fut_expiration_date,
e.trade_date,
timestamp_table.timeformated,
a.rate * a.basis_value AS dma_fee,
SUM(A.RATE * A.BASIS_VALUE) OVER () as DMA_FEE_TOTAL
FROM
repldoadm.ire_estimate_trans_map a
LEFT JOIN repldoadm.ire_charges_estimate b ON a.estimate_id = b.estimate_id
LEFT JOIN repldoadm.ire_charges_lu c ON b.charge_id = c.charge_id
LEFT JOIN repldoadm.ire_charge_types_lu d ON c.charge_type_id = d.charge_type_id
LEFT JOIN repldoadm.vw_ire_trade_transactions e ON a.transaction_id = e.trans_id
LEFT JOIN (
SELECT
to_char(create_ts, 'HH24:MI:SS') AS timeformated,
trans_id
FROM
repldoadm.vw_ire_trade_transactions
) timestamp_table ON ( a.transaction_id = timestamp_table.trans_id )
LEFT JOIN repldoadm.vw_ire_accounts f ON e.account_name = f.pb_account_name
LEFT JOIN stig_adm.instrument_universe g ON e.tradeable_instr_name = g.short_name
WHERE
timeformated >= '07:00:00'
AND timeformated <= '18:00:00'
AND fut_expiration_date >= add_months(trunc(sysdate), - 12)
AND trade_date = '02-AUG-21'
AND e.exchange_id = '30'
AND charge_group = 'EXECUTION_BROKER_CHARGES'
AND estimate_run = 'Late'
--AND transaction_id IN ( '791088074', '791309067' )
)
group by TRADEABLE_INSTR_NAME, TRADING_USERID, UNDERLYING_SYM_BLOOMBERG;

Answer 1

Group by TRUNC(trade_date, 'MM')
SELECT TRADEABLE_INSTR_NAME,
TRADING_USERID,
UNDERLYING_SYM_BLOOMBERG,
SUM(DMA_FEE) AS DMA_FEE_SUBTOTAL,
TRUNC( trade_date, 'MM' ) AS trade_month
FROM (
<your huge sub-query>
)
group by
TRADEABLE_INSTR_NAME,
TRADING_USERID,
UNDERLYING_SYM_BLOOMBERG,
TRUNC(trade_date, 'MM');
Your query would effectively be:
select TRADEABLE_INSTR_NAME,
TRADING_USERID,
UNDERLYING_SYM_BLOOMBERG,
TRUNC(trade_date, 'MM') AS trade_month
SUM(DMA_FEE) AS DMA_FEE_SUBTOTAL
FROM (
SELECT e.trading_userid,
e.tradeable_instr_name,
g.underlying_sym_bloomberg,
e.trade_date,
a.rate * a.basis_value AS dma_fee
FROM repldoadm.ire_estimate_trans_map a
INNER JOIN repldoadm.ire_charges_estimate b
ON a.estimate_id = b.estimate_id
INNER JOIN repldoadm.ire_charges_lu c
ON b.charge_id = c.charge_id
INNER JOIN repldoadm.ire_charge_types_lu d
ON c.charge_type_id = d.charge_type_id
INNER JOIN repldoadm.vw_ire_trade_transactions e
ON ( a.transaction_id = e.trans_id
AND TO_CHAR(e.create_ts, 'HH24:MI:SS') BETWEEN '07:00:00'
AND '18:00:00'
)
LEFT JOIN repldoadm.vw_ire_accounts f
ON e.account_name = f.pb_account_name
INNER JOIN stig_adm.instrument_universe g
ON e.tradeable_instr_name = g.short_name
WHERE g.fut_expiration_date >= add_months(trunc(sysdate), - 12)
AND e.trade_date = DATE '2021-08-02'
AND e.exchange_id = '30'
AND d.charge_group = 'EXECUTION_BROKER_CHARGES'
AND estimate_run = 'Late'
--AND transaction_id IN ( '791088074', '791309067' )
)
group by
TRADEABLE_INSTR_NAME,
TRADING_USERID,
UNDERLYING_SYM_BLOOMBERG,
TRUNC(trade_date, 'MM');
Note: Because you are requiring non-NULL values in the WHERE clause your LEFT JOINs become INNER JOINs.
Note: You do not need to SELECT columns in the inner query that you are not using in the outer query.
Note: I do not know if you need to join repldoadm.vw_ire_accounts f as you do not appear to use it.