I have this query that I am working on for a one time report. Basically what I am trying to do is find all the records where there are more than two transactions by a single account ID last month. I know it is probably something easy, my mind is just blanking.
SELECT streaming_transactions.account_id,
streaming_transactions_detail.transactions_description,
streaming_transactions_detail.transactions_detail_id,
streaming_transactions_detail.transactions_id,
streaming_transactions_detail.transactions_detail_amount,
streaming_transactions_detail.detail_type,
streaming_transactions_detail.products_id,
streaming_transactions_detail.products_levels_id,
streaming_transactions_detail.subscriptions_id,
streaming_transactions_detail.subscriptions_payment_options_id,
streaming_transactions_detail.modified
FROM streaming_transactions_detail
INNER JOIN streaming_transactions ON streaming_transactions_detail.transactions_id = streaming_transactions.transactions_id
WHERE streaming_transactions.charged = 1
AND streaming_transactions.timestamp_inserted > '2009-09-01 00:00:00'
AND streaming_transactions.account_id IN (
SELECT account_id
FROM streaming_transactions_detail
WHERE modified > '2009-09-01 00:00:00'
AND count(account_id) > 1)
AND streaming_transactions_detail.transactions_description LIKE '%Service Subscription%'
ORDER BY streaming_transactions.account_id DESC
I think you're almost there.
The subquery to get the transaction Ids is off, however
SELECT account_id
FROM streaming_transactions_detail
WHERE modified > '2009-09-01 00:00:00'
AND count(account_id) > 1)
-- Should be something like
SELECT account_id, COUNT(account_id)
FROM streaming_transactions_detail
WHERE modified > '2009-09-01 00:00:00'
GROUP BY account_id
HAVING count(account_id) > 1)
[unrelated] I'll throw in an unsolicited hint about style.
By using table aliases, you can improve the readabilty of the query. This can be done by optionally adding "AS xyz" where xyz is some short but mnemonic name, unique to this query, you can use xyz wherever in the query where you would use the long_named_table.
For example:
FROM streaming_transactions_detail AS D
and then
SELECT streaming_transactions.account_id,
streaming_transactions_detail.transactions_description,
streaming_transactions_detail.transactions_detail_id,
...
Can become (optionally, i.e. the "streaming_transactions_detail." still works too)
SELECT D.account_id,
D.transactions_description,
D.transactions_detail_id,
...
Related
Thanks for the help on this matter, I'm new with SQL. I'm trying to get a sub-count of Jedi who had more than 2 padawans last month. I tried putting the condition in WHERE but I get an error saying I can't include aggregates in it. I also tried using a CASE but kept getting a syntax error there too. Any help on this would be incredible. Thank you so much!
SELECT COUNT(DISTINCT old_republic.jedi_id), old_republic.region_id
FROM jedi_archives.old_repulicdata old_republic
WHERE old_republic.republic_date >= '2022-06-01' AND old_republic.republic_date <= '2022-06-30' AND COUNT(old_republic.padawan)>2
GROUP BY old_republic.region_id
ORDER BY old_republic.region_id
SELECT old_republic.jedi_id CASE (
WHEN Count(old_republic.padawan)>2
THEN 1
ELSE 0 End), old_republic.region_id
FROM jedi_archives.old_repulicdata old_republic
WHERE old_republic.republic_date >= '2022-06-01' AND old_republic.republic_date <= '2022-06-30'
GROUP BY old_republic.region_id
ORDER BY old_republic.region_id
I can't comment to ask for a fiddle, but from what you've written, you're probably looking for the HAVING clause.
Assuming that padawan denotes the number of Padawans:
SELECT region_id, jedi_id, sum(padawan)
FROM jedi_archives.old_republicdata
WHERE republic_date >= '2022-06-01'
AND republic_date <= '2022-06-30'
GROUP BY region_id, jedi_id
HAVING sum(padawan) > 2;
This query will return the sum of Padawans for each Jedi per region who had more than two Padawans last month in one region (if you don't want to take the region into account, remove it from the SELECT and GROUP BY clause). Other Jedis won't appear in the result.
You can use the CASE expression, too, in order to indicate whether a Jedi had more than two padawans:
SELECT region_id, jedi_id,
CASE WHEN sum(padawan) > 2 THEN 1 ELSE 0 END AS more_than_2_padawans
FROM jedi_archives.old_republicdata
WHERE republic_date >= '2022-06-01'
AND republic_date <= '2022-06-30'
GROUP BY region_id, jedi_id;
I'm not entirely sure without sample data. But I think using the HAVING clause could solve your question.
SELECT COUNT(jedi_id) as jedi_id, region_id FROM tableA
WHERE republic_date between '2022-05-20' and '2022-05-25'
GROUP BY region_id
having padawan > 2
db fiddle
Right now I have a query that lets me know which users didn't make a purchase 12 months prior to becoming members. These users have MEM_PRE_12=0 and I want to filter off those users more natively using SQL partitions rather than always putting rudimentary WHERE criteria.
Here is the SQL I use to find the users I want/don't want.
SELECT SUM(CASE WHEN DATE <= DATEADD(month, -12, U.INSERTED_AT) THEN 1 ELSE 0 END) AS MEM_PRE_12, I.CLIENTID, I.INSTALLATIONID
FROM <<<My_Joined_Tables>>>
GROUP BY I.CLIENTID, I.INSTALLATIONID
HAVING MEM_PRE_12 != 0
ORDER BY MEM_PRE_12
After this I'm going to have to go back and say where I.CLIENTID in the above nested query and select the actual information I want from users who made purchases greater than their insertion date.
How can I do this without so much nesting of all these joined tables?
If you want the detailed rows for customers who made a purchase in the last 12 months, you can use window functions:
with q as (
<whatever your query logic is>
)
select q.*
from (select q.*,
SUM(CASE WHEN DATE <= DATEADD(month, -12, U.INSERTED_AT) THEN 1 ELSE 0 END) over (partition by CLIENTID, INSTALLATIONID) as AS MEM_PRE_12
from q
) q
where mem_pre_12 > 0;
I am trying to run a query on an Oracle 10g DB to try and view 2 groups of transactions. I want to view basically anyone who has a transaction this year (2014) that also had a transaction in the previous 5 years. I then want to run a query for anyone who has a transaction this year (2014) that hasn't ordered from us in the last 5 years. I assumed I could do this with the 'IN' and 'NOT IN' features. The 'IN' query runs fine but the 'NOT IN' never completes. DB is fairly large which is probably why. Would love any suggestions from the experts!
*Notes, [TEXT] is a description of our Customer's Company name, sometimes the accounting department didn't tie this to our customer ID which left NULL values, so using TEXT as my primary grouping seemed to work although the name is obscure. CODE_D is a product line just to bring context to the name.
Below is my code:
SELECT CODE_D, sum(coalesce(credit_amount, 0) - coalesce(debet_amount,0)) as TOTAL
FROM
gen_led_voucher_row_tab
WHERE ACCOUNTING_YEAR like '2014'
and TEXT NOT IN
(select TEXT
from gen_led_voucher_row_tab
and voucher_date >= '01-JUN-09'
and voucher_date < '01-JUN-14'
and (credit_amount > '1' or debet_amount > '1')
)
GROUP BY CODE_D
ORDER BY TOTAL DESC
Try using a LEFT JOIN instead of NOT IN:
SELECT t1.CODE_D, sum(coalesce(t1.credit_amount, 0) - coalesce(t1.debet_amount,0)) as TOTAL
FROM gen_led_voucher_row_tab AS t1
LEFT JOIN gen_led_voucher_row_tab AS t2
ON t1.TEXT = t2.TEXT
AND t2.voucher_date >= '01-JUN-09'
AND t2.voucher_date < '01-JUN-14'
AND (credit_amount > '1' or debet_amount > '1')
WHERE t2.TEXT IS NULL
AND t1.ACCOUNTING_YEAR = '2014'
GROUP BY CODE_D
ORDER BY TOTAL DESC
ALso, make sure you have an index on the TEXT column.
You can increase your performance by changing the Not In clause to a Where Not Exists like as follows:
Where Not Exists
(
Select 1
From gen_led_voucher_row_tab b
Where voucher_date >= '01-JUN-09'
and voucher_date < '01-JUN-14'
and (credit_amount > '1' or debet_amount > '1')
And a.Text = b.Text
)
You'll need to alias the first table as well to a for this to work. Essentially, you're pulling back a ton of data to just discard it. Exists invokes a Semi Join which does not pull back any data at all, so you should see significant improvement.
Edit
Your query, as of the current update to the question should be this:
SELECT CODE_D,
sum(coalesce(credit_amount, 0) - coalesce(debet_amount,0)) as TOTAL
FROM gen_led_voucher_row_tab a
Where ACCOUNTING_YEAR like '2014'
And Not Exists
(
Select 1
From gen_led_voucher_row_tab b
Where voucher_date >= '01-JUN-09'
and voucher_date < '01-JUN-14'
and (credit_amount > '1' or debet_amount > '1')
And a.Text = b.Text
)
GROUP BY CODE_D
ORDER BY TOTAL DESC
In SQL Server 2008, this query works:
SELECT
SUM(CAST(isredeemed AS TINYINT)) AS totalredeemed,
FROM rewards
GROUP BY merchantid
It gives you the number of redeemed rewards by merchant. The TINYINT cast is need to avoid the error Operand data type bit is invalid for sum operator.
Now I'd like to do a similar query, but one that only finds rewards redeemed in the last few days. I tried this:
SELECT
SUM(CAST((isredeemed & ( MIN(dateredeemed) > '2014-01-10 05:00:00')) AS TINYINT)) AS claimedthisweek,
FROM rewards
GROUP BY merchantid
and I get the error
Incorrect syntax near '>'.
I also tried replacing & with && and also with AND. But those don't work either.
How can I make the example work?
This question lacks detail to give an exact answer but you need to use a derived table or subquery for the calculation.
something like this
SELECT r1.merchantid, r2.claimedthisweek
FROM rewards r1
JOIN (
SELECT merchantid, SUM(CAST(isredeemed AS INT)) claimedthisweek
FROM rewards
GROUP BY merchantid
HAVING MIN(dateredeemed) > '20140101'
) r2
This should be on the having clause, like this:
SELECT
SUM(CAST((isredeemed AS TINYINT)) AS claimedthisweek
FROM rewards
GROUP BY merchantid
HAVING MIN(dateredeemed) > '2014-01-10 05:00:00'
Any reason not to do the filtering at the "where" clause level? That should work so long as all rows you're aggregating match the same criteria:
SELECT
SUM(CAST(isredeemed AS TINYINT)) AS claimedthisweek
FROM rewards
WHERE dateredeemed > '2014-01-10 05:00:00'
GROUP BY merchantid
I think you want this would work but it would count some twice:
SELECT
SUM(case when MIN(dateredeemed) > '2014-01-10 05:00:00' then 1 else 0 end )
FROM rewards
GROUP BY merchantid
My initial query looks like this:
select process_date, count(*) batchCount
from T1.log_comments
order by process_date asc;
I need to be able to do some quick analysis for weekends that are missing, but wanted to know if there was a quick way to fill in the missing dates not present in process_date.
I've seen the solution here but am curious if there's any magic hidden in db2 that could do this with only a minor modification to my original query.
Note: Not tested, framed it based on my exposure to SQL Server/Oracle. I guess this gives you the idea though:
*now amended and tested on DB2*
WITH MaxDateQry(MaxDate) AS
(
SELECT MAX(process_date) FROM T1.log_comments
),
MinDateQry(MinDate) AS
(
SELECT MIN(process_date) FROM T1.log_comments
),
DatesData(ProcessDate) AS
(
SELECT MinDate from MinDateQry
UNION ALL
SELECT (ProcessDate + 1 DAY) FROM DatesData WHERE ProcessDate < (SELECT MaxDate FROM MaxDateQry)
)
SELECT a.ProcessDate, b.batchCount
FROM DatesData a LEFT JOIN
(
SELECT process_date, COUNT(*) batchCount
FROM T1.log_comments
) b
ON a.ProcessDate = b.process_date
ORDER BY a.ProcessDate ASC;