I am using below query to calculate business days between two dates for all the order numbers. Business days are already available in the teradata table Common_WorkingCalendar. But, i'm also facing spool space issue while i execute the query. I have ample space available in my data lab. Need to optimize the query. Appreciate any inputs.
SELECT
tx."OrderNumber",
(SELECT COUNT(1) FROM Common_WorkingCalendar
WHERE CalDate between Cast(tx."TimeStamp" as date) and Cast(mf.ShipDate as date)) as BusDays
from StoreFulfillment ff
inner join StoreTransmission tx
on tx.OrderNumber = ff.OrderNumber
inner join StoreMerchandiseFulfillment mf
on mf.OrderNumber = ff.OrderNumber
This is a very inefficient way to get this count which results in a product join.
The recommended approach is adding a sequential number to your calendar which increases only on business days (calculated using SUM(CASE WHEN businessDay THEN 1 ELSE 0 END) OVER (ORDER BY CalDate ROWS UNBOUNDED PRECEDING)), then it's two joins, for the start date and the end date.
If this calculation is needed a lot you better add a new column, otherwise you can do it on the fly:
WITH cte AS
(
SELECT CalDate,
-- as this table only contains business days you can use this instead
row_number(*) Over (ORDER BY CalDate) AS DayNo
FROM Common_WorkingCalendar
)
SELECT
tx."OrderNumber",
to_dt.DayNo - from_dt.DayNo AS BusDays
FROM StoreFulfillment ff
INNER JOIN StoreTransmission tx
ON tx.OrderNumber = ff.OrderNumber
INNER JOIN StoreMerchandiseFulfillment mf
ON mf.OrderNumber = ff.OrderNumber
JOIN cte AS from_dt
ON from_dt.CalDate = Cast(tx."TimeStamp" AS DATE)
JOIN cte AS to_dt
ON to_dt.CalDate = Cast(mf.ShipDate AS DATE)
Related
I have a huge table with over 1 million transaction records and I need to join this table to itself and pull all similar transactions within 52 weeks prior for each transaction and aggregate them for later use in an ML model.
select distinct a.transref,
a.transdate, a.transamount,
a.transtype,
avg (b.transamount)
over (partition by a.transref,a.transdate, a.transamount,a.transtype) as avg_trans_amount
from trans_table a
inner join trans_table b
on a.transtype = b.transtype
and b.transdate <= dateadd(week, -52, a.transdate)
and b.transdate <= a.transdate
and a.transdate between '2022-11-16' and '2021-11-16'
the transaction table looks like this:
+--------+----------+-----------+---------+
|transref|trasndate |transamount|transtype|
+--------+----------+-----------+---------+
|xh123rdk|2022-11-16|112.48 |food & Re|
|g8jegf90|2022-11-04|23.79 |Misc |
|ulpef32p|2022-10-23|83.15 |gasoline |
+--------+----------+-----------+---------+
and the expected output should look like this:
+--------+----------+-----------+---------+----------------+
|transref|trasndate |transamount|transtype|avg_trans_amount|
+--------+----------+-----------+---------+----------------+
|xh123rdk|2022-11-16|112.48 |food & Re|180.11 |
|g8jegf90|2022-11-04|23.79 |Misc |43.03 |
|ulpef32p|2022-10-23|83.15 |gasoline |112.62 |
+--------+----------+-----------+---------+----------------+
Since each transaction may pull over 10,000 similar type records the query is very slow and expensive to run, therefore SQL Server failed to create the output table.
How can I optimize this query to run efficiently within a reasonable time?
Note: After failing to run the query, I ended up creating a stored procedure to split the original table a into smaller chunks, join it to the big table, aggregate the results and append the results to an output table and repeat this until the entire table a was covered. This way I could manage to do the job, however, it was still slow. I expect there are better ways to do it in SQL without all this manual work.
ok, I think I figured out what's causing the query to run tooslow. the trick is to avoid repetitive and unnecessary calculations by doing some group by first before doing the join.
with merch as (
select transtype,
dateadd(week, -52, transdate) as startdate,
transdate as enddate),
from trans_table
group by transtype, transdate),
summary as (
select distinct transtype,
stratdate, enddate,
avg(t.transamt) over (partition by
m.transtype, m.startdate, m.enddate) as avg_amt,
percentile_cont(0.5) within group (order by t.transamt) over (partition by
m.transtype, m.startdate, m.enddate) as median_amt
from merch as m
inner join trans_table as t
on m.transtype = t.transdate and
t.transdate between m.starttype and
m.enddate)
select t.*, s.avg_amt s.median_amt
from trans_table t
inner join summary s
on t.transtype = s.transtype
and t.transdate = s.enddate
I have a query that's intended as the base dataset for an AR Aging report in a BI tool. The report has to be able to show AR as of a given date across a several-month range. I have the logic working, but I'm seeing pretty slow performance. Code below:
WITH
DAT AS (
SELECT
MY_DATE AS_OF_DATE
FROM
NS_REPORTS."PUBLIC".NETSUITE_DATE_TABLE
WHERE
CAST(CAST(MY_DATE AS TIMESTAMP) AS DATE) BETWEEN '2020-01-01' AND CAST(CAST(CURRENT_DATE() AS TIMESTAMP) AS DATE)
), INV AS
(
WITH BASE AS
(
SELECT
BAS1.TRANSACTION_ID
, DAT.AS_OF_DATE
, SUM(BAS1.AMOUNT) ORIG_AMOUNT_BASE
FROM
"PUBLIC".BILL_TRANS_LINES_BASE BAS1
CROSS JOIN DAT
WHERE
BAS1.TRANSACTION_TYPE = 'Invoice'
AND BAS1.TRANSACTION_DATE <= DAT.AS_OF_DATE
--AND BAS1.TRANSACTION_ID = 6114380
GROUP BY
BAS1.TRANSACTION_ID
, DAT.AS_OF_DATE
)
, TAX AS
(
SELECT
TRL1.TRANSACTION_ID
, SUM(TRL1.AMOUNT_TAXED * - 1) ORIG_AMOUNT_TAX
FROM
CONNECTORS.NETSUITE.TRANSACTION_LINES TRL1
WHERE
TRL1.AMOUNT_TAXED IS NOT NULL
AND TRL1.TRANSACTION_ID IN (SELECT TRANSACTION_ID FROM BASE)
GROUP BY
TRL1.TRANSACTION_ID
)
SELECT
BASE.TRANSACTION_ID
, BASE.AS_OF_DATE
, BASE.ORIG_AMOUNT_BASE
, COALESCE(TAX.ORIG_AMOUNT_TAX, 0) ORIG_AMOUNT_TAX
FROM
BASE
LEFT JOIN TAX ON TAX.TRANSACTION_ID = BASE.TRANSACTION_ID
)
SELECT
AR.*
, CASE
WHEN AR.DAYS_OUTSTANDING < 0
THEN 'Current'
WHEN AR.DAYS_OUTSTANDING BETWEEN 0 AND 30
THEN '0 - 30'
WHEN AR.DAYS_OUTSTANDING BETWEEN 31 AND 60
THEN '31 - 60'
WHEN AR.DAYS_OUTSTANDING BETWEEN 61 AND 90
THEN '61 - 90'
WHEN AR.DAYS_OUTSTANDING > 90
THEN '91+'
ELSE NULL
END DO_BUCKET
FROM
(
SELECT
AR1.*
, TRA1.TRANSACTION_TYPE
, DATEDIFF('day', AR1.AS_OF_DATE, CAST(CAST(TRA1.DUE_DATE AS TIMESTAMP) AS DATE)) DAYS_OUTSTANDING
, AR1.ORIG_AMOUNT_BASE + AR1.ORIG_AMOUNT_TAX + AR1.PMT_AMOUNT AMOUNT_OUTSTANDING
FROM
(
SELECT
INV.TRANSACTION_ID
, INV.AS_OF_DATE
, INV.ORIG_AMOUNT_BASE
, INV.ORIG_AMOUNT_TAX
, COALESCE(PMT.PMT_AMOUNT, 0) PMT_AMOUNT
FROM
INV
LEFT JOIN (
SELECT
TLK.ORIGINAL_TRANSACTION_ID
, DAT.AS_OF_DATE
, SUM(TLK.AMOUNT_LINKED * - 1) PMT_AMOUNT
FROM
CONNECTORS.NETSUITE."TRANSACTION_LINKS" AS TLK
CROSS JOIN DAT
WHERE
TLK.LINK_TYPE = 'Payment'
AND CAST(CAST(TLK.ORIGINAL_DATE_POSTED AS TIMESTAMP) AS DATE) <= DAT.AS_OF_DATE
GROUP BY
TLK.ORIGINAL_TRANSACTION_ID
, DAT.AS_OF_DATE
) PMT ON PMT.ORIGINAL_TRANSACTION_ID = INV.TRANSACTION_ID
AND PMT.AS_OF_DATE = INV.AS_OF_DATE
) AR1
JOIN CONNECTORS.NETSUITE."TRANSACTIONS" TRA1 ON TRA1.TRANSACTION_ID = AR1.TRANSACTION_ID
)
AR
WHERE
1 = 1
--AND CAST(AMOUNT_OUTSTANDING AS NUMERIC(15, 2)) > 0
AND AS_OF_DATE >= '2020-04-22'
As you can see, I'm using a date table for the as-of date logic. I think this is the best way to do it, but I welcome any suggestions for better practice.
If I run the query with a single as-of date, it takes 1 min 6 sec and the two main aggregates, on TRANSACTION_LINKS and BILL_TRANS_LINES_BASE, each take about 25% of processing time. I'm not sure why. If I run with the filter shown, >= '2020-04-22', it takes 3 min 33 sec and the aggregates each take about 10% of processing time; they're lower because the ResultWorker takes 63% of processing time to write the results because it's so many rows.
I'm new to Snowflake but not to SQL. My understanding is that Snowflake does not allow manual creation of indexes, but again, I'm happy to be wrong. Please let me know if you have any ideas for improving the performance of this query.
Thanks in advance.
EDIT 1:
Screenshot of most expensive node in query profile
Without seeing the full explain plan and having some sample data to play with it is difficult to give any definitive answers, but here a few thoughts, for what they are worth...
The first are more about readability and may not help performance much:
Don't embed CTEs within each other, just define them in the order that they are needed. There is no need to define BASE and TAX within INV
Use CTEs as much as possible. Your main SELECT statement has 2 other SELECT statements embedded within it. It would be much more readable if these were defined using CTEs
Specific performance issues:
Keep data volumes as low as possible for as long as possible. Your CROSS JOINs obviously create cartesian products that massively increases the volume of data - therefore implement this as late in your SQL as possible rather than right at the start as you have done
While it may make your SQL less readable, use as few SQL statements as possible. For example, you should be able to create your INV CTE with a single SELECT statement rather than the 3 statements/CTEs that you are using
I currently have a query running to average survey scores for agents. We use the date range of the LastDayOfTheQuarter and 180 days back to calculate these scores. I ran into an issue for this current quarter.
One of my agents hasn't received any surveys in 2020 which is causing the query to not pull the current lastdayofquarter and 180 days back of results.
The code I am using:
SELECT
Agent,
U.Position,
U.BranchDescription,
(ADDDATE(LastDayOfQuarter, -180)) AS MinDate,
(LastDayOfQuarter) AS MaxDate,
COUNT(DISTINCT Response ID) as SurveyCount,
AVG(CASE WHEN Question ID = Q1_2 THEN Answer Value END) AS EngagedScore,
AVG(CASE WHEN Question ID = Q1_3 THEN Answer Value END) AS KnowledgableScore,
AVG(CASE WHEN Question ID = Q1_6 THEN Answer Value END) AS ValuedScore
FROM qualtrics_responses
LEFT JOIN date D
ON (D.`Date`) = (DATE(`End Date`))
LEFT JOIN `users` U
ON U.`UserID` = `Agent ID`
WHERE `Agent` IS NOT NULL
AND DATE(`End Date`) <= (`LastDayOfQuarter`)
AND DATE(`End Date`) >= (ADDDATE(`LastDayOfQuarter`, -180))
GROUP BY `Agent`, (ADDDATE(`LastDayOfQuarter`, -180))
i know the issue is due to the way I am joining the dates and since he doesn't have a result in this current year, the end date to date join isn't grabbing the desired date range. I can't seem to come up with any alternatives. Any help is appreciated.
I make the assumption that table date in your query is a calendar table, that stores the starts and ends of the quarters (most likely with one row per date in the quarter).
If so, you can solve this problem by rearranging the joins: first cross join the users and the calendar table to generate all possible combinations, then bring in the surveys table with a left join:
SELECT
U.UserID,
U.Position,
U.BranchDescription,
D.LastDayOfQuarter - interval 180 day AS MinDate,
D.LastDayOfQuarter AS MaxDate,
COUNT(DISTINCT Q.ResponseID) as SurveyCount,
AVG(CASE WHEN Q.QuestionID = 'Q1_2' THEN Q.Answer Value END) AS EngagedScore,
AVG(CASE WHEN Q.QuestionID = 'Q1_3' THEN Q.Answer Value END) AS KnowledgableScore,
AVG(CASE WHEN Q.QuestionID = 'Q1_6' THEN Q.Answer Value END) AS ValuedScore
FROM date D
CROSS JOIN users U
LEFT JOIN qualtrics_responses Q
ON Q.EndDate >= D.Date
AND Q.EndDate < D.Date + interval 1 day
AND U.UserID = Q.AgentID
AND Q.Agent IS NOT NULL
GROUP BY
U.UserID,
U.Position,
U.BranchDescription,
D.LastDayOfQuarter
Notes:
I adapted the date arithmetics - this assumes that you are using MySQL, as the syntax of the query suggests
You should really qualify all the columns in the query, by prefixing them with the alias of the table they belong to; this makes the query so much easier to understand. I gave a tried at it, you might need to review that.
All non-aggregated columns should appear in the group by clause (also see the comment from Eric); this is a a requirement in most databaseses, and good practice anywhere
I'm having a bit of an issue wrapping my head around the logic of this changing dimension. I would like to associate these two tables below. I need to match the Cost - Period fact table to the cost dimension based on the Id and the effective date.
As you can see - if the month and year field is greater than the effective date of its associated Cost dimension, it should adopt that value. Once a new Effective Date is entered into the dimension, it should use that value for any period greater than said date going forward.
EDIT: I apologize for the lack of detail but the Cost Dimension will actually have a unique Index value and the changing fields to reference for the matching would be Resource, Project, Cost. I tried to match the query you provided with my fields, but I'm getting the incorrect output.
FYI: Naming convention change: EngagementId is Id, Resource is ConsultantId, and Project is ProjectId
I've changed the images below and here is my query
,_cte(HoursWorked, HoursBilled, Month, Year, EngagementId, ConsultantId, ConsultantName, ProjectId, ProjectName, ProjectRetainer, RoleId, Role, Rate, ConsultantRetainer, Salary, amount, EffectiveDate)
as
(
select sum(t.Duration), 0, Month(t.StartDate), Year(t.StartDate), t.EngagementId, c.ConsultantId, c.ConsultantName, c.ProjectId, c.ProjectName, c.ProjectRetainer, c.RoleId, c.Role, c.Rate, c.ConsultantRetainer,
c.Salary, 0, c.EffectiveDate
from timesheet t
left join Engagement c on t.EngagementId = c.EngagementId and Month(c.EffectiveDate) = Month(t.EndDate) and Year(c.EffectiveDate) = Year(t.EndDate)
group by Month(t.StartDate), Year(t.StartDate), t.EngagementId, c.ConsultantName, c.ConsultantId, c.ProjectId, c.ProjectName, c.ProjectRetainer, c.RoleId, c.Role, c.Rate, c.ConsultantRetainer,
c.Salary, c.EffectiveDate
)
select * from _cte where EffectiveDate is not null
union
select _cte.HoursWorked, _cte.HoursBilled, _cte.Month, _cte.Year, _cte.EngagementId, _cte.ConsultantId, _cte.ConsultantName, _cte.ProjectId, _Cte.ProjectName, _cte.ProjectRetainer, _cte.RoleId, _cte.Role, sub.Rate, _cte.ConsultantRetainer,_cte.Salary, _cte.amount, sub.EffectiveDate
from _cte
outer apply (
select top 1 EffectiveDate, Rate
from Engagement e
where e.ConsultantId = _cte.ConsultantId and e.ProjectId = _cte.ProjectId and e.RoleId = _cte.RoleId
and Month(e.EffectiveDate) < _cte.Month and Year(e.EffectiveDate) < _cte.Year
order by EffectiveDate desc
) sub
where _cte.EffectiveDate is null
Example:
I'm struggling with writing the query that goes along with this. At first I attempted to partition by greatest date. However, when I executed the join I got the highest effective date for every single period (even those prior to the effective date).
Is this something that can be accomplished in a query or should I be focusing on incremental updates of the destination table so that any effective date / time period in the past is left alone?
Any tips would be great!
Thanks,
Channing
Try this one:
; with _CTE as(
select p.* , c.EffectiveDate, c.Cost
from period p
left join CostDimension c on p.id = c.id and p.Month = DATEPART(month, c.EffectiveDate) and p.year = DATEPART (year, EffectiveDate)
)
select * from _CTE Where EffectiveDate is not null
Union
select _CTE.id, _CTE.Month, _CTE.Year, sub.EffectiveDate, sub.Cost
from _CTE
outer apply (select top 1 EffectiveDate, Cost
from CostDimension as cd
where cd.Id = _CTE.id and cd.EffectiveDate < DATETIMEFROMPARTS(_CTE.Year, _CTE.Month, 1, 0, 0, 0, 0)
order by EffectiveDate desc
) sub
where _Cte.EffectiveDate is null
Im trying to generate a running total by month and year. Ive tried a few examples but I cant get it working. This is the sql I have and I would want to create a running total for the totalclients column
Month| Year| TotalClients| Running Total
Jan |2014| 1| 1
Feb| 2014| 4| 5
Mar| 2014| 8| 13
select Month, Year, TotalClients
From Total
This was the code I was trying to use, ive used a declare table as the main data comes from a different query but this should be the bit you need. I also commented out one of the from lines as I was trying out both way, the commented out line was in a few examples on the net but I couldn't get it working
select t1.monthstart, t1.yearstart, t1.TotalClients, sum(t2.TotalClients) as 'RunningTotal'
from #Totals t1 inner join #Totals t2 on t1.monthstart = t2.monthstart and t1.yearstart = t2.yearstart
--from #Totals t1, #Totals t2
WHERE t1.MonthStart <= t2.MonthStart and t1.Yearstart <= t2.Yearstart
GROUP BY t1.Yearstart, t1.MonthStart, t1.TotalClients
ORDER BY t1.yearstart , t1.monthstart
As #xQbert posted in comments above (I advise reading that article), SQL Server "Windowing Functions" is what you want to use in version 2012+. Windowing functions are flexible and powerful, and far more efficient than self-joins.
As an actual answer, here would be some possible code for you to use:
SELECT YearStart, MonthStart,
ClientCount = SUM(TotalClients) OVER (
PARTITION BY YearStart, MonthStart
ORDER BY YearStart, MonthStart RANGE UNBOUNDED PRECEDING
)
FROM Totals t1
ORDER BY YearStart, MonthStart
I used this in the end, I added a faulldate in to simplify what I wanted and it worked, I think the issue was in the join I used it had the <= the wrong way around.
SELECT
st1.invoicestartdate,
st1.TotalClients,
RunningTotal = SUM(st2.TotalClients)
FROM
#Totals AS st1
INNER JOIN
#Totals AS st2
ON st2.invoicestartdate <= st1.invoicestartdate
GROUP BY st1.invoicestartdate, st1.TotalClients
ORDER BY st1.invoicestartdate;
This query works for SQL Server 2012 and up. I assumed Month is numeric (Jan = 1, Feb = 2, etc.)
SELECT *,
SUM(t.TotalClients) OVER (PARTITION BY t.[Year] ORDER BY t.[Month])
FROM #Totals t
It will reset the client count once the year changes. To keep it going, change the SUM clause to
SUM(t.TotalClients) OVER (ORDER BY t.[Year], t.[Month])