Oracle tables
The following query works as START_TIME_ and DT.LABEL_DATE_TIME_15MIN_INT output is like for like. I have to admit it's the first time I've written a join like this, is it acceptable or is there a more efficient/proper way of accomplishing this?
select
A.ID,
trunc(A.START_TIME,'dd') + (round(to_char(A.START_TIME,'sssss') / 900) / 96) as START_TIME_,
DT.LABEL_DATE_TIME_15MIN_INT
from TBL_A A
inner join DATE_TIME DT
on ((trunc(A.IN_TIME_START,'dd') + (round(to_char(A.IN_TIME_START,'sssss') / 900) / 96)) = DT.DATE_TIME_15MIN_INT)
order by A.ID;
Reason for approach: A.START_TIME is TIMESTAMP(6) GMT-5 DT table has EPOCH time Keys in GMT with corresponding "Formatted Labels" adjusted to GMT-5. Instead of converting A.START_TIME to EPOCH, adjusting for timezone, and joining on the key I performed the arithmetic to A to the correct 15min interval then joined to the label. Am I missing a really simple solution here?
I would go for a more concise approach:
SELECT
a.id,
a.start_time_,
dt.label_date_time_15min_int
FROM
date_time dt
JOIN (
SELECT
id,
TRUNC(in_time_start,'dd') + (ROUND(TO_CHAR(in_time_start,'sssss') / 900) / 96) AS in_time_start_,
TRUNC(start_time,'dd') + (ROUND(TO_CHAR(start_time,'sssss') / 900) / 96) AS start_time_
FROM
tbl_a
) a ON a.in_time_start_ = dt.date_time_15min_int
ORDER BY
a.id
;
But fundamentally there's not much difference. Your join is valid, and you could also put it in the where clause if you wanted to. At this point there's not much you can do to optimize it without structural and applicative changes.
Related
I'm wondering if it's possible to pass one or more parameters to a WITH clause query; in a very simple way, doing something like this (taht, obviously, is not working!):
with qq(a) as (
select a+1 as increment
from dual
)
select qq.increment
from qq(10); -- should get 11
Of course, the use I'm going to do is much more complicated, since the with clause should be in a subquery, and the parameter I'd pass are values taken from the main query....details upon request... ;-)
Thanks for any hint
OK.....here's the whole deal:
select appu.* from
(<quite a complex query here>) appu
where not exists
(select 1
from dual
where appu.ORA_APP IN
(select slot from
(select distinct slots.inizio,slots.fine from
(
with
params as (select 1900 fine from dual)
--params as (select app.ora_fine_attivita fine
-- where app.cod_agenda = appu.AGE
-- and app.ora_fine_attivita = appu.fine_fascia
--and app.data_appuntamento = appu.dataapp
--)
,
Intervals (inizio, EDM) as
( select 1700, 20 from dual
union all
select inizio+EDM, EDM from Intervals join params on
(inizio <= fine)
)
select * from Intervals join params on (inizio <= fine)
) slots
) slots
where slots.slot <= slots.fine
)
order by 1,2,3;
Without going in too deep details, the where condition should remove those records where 'appu.ORA_APP' match one of the records that are supposed to be created in the (outer) 'slots' table.
The constants used in the example are good for a subset of records (a single 'appu.AGE' value), that's why I should parametrize it, in order to use the commented 'params' table (to be replicated, then, in the 'Intervals' table.
I know thats not simple to analyze from scratch, but I tried to make it as clear as possible; feel free to ask for a numeric example if needed....
Thanks
I need to query 2 tables, one contains a TIMESTAMP(6) column, other contains a DATE column. I want to write a select statement that prints both values and diff between these two in third column.
SB_BATCH.B_CREATE_DT - timestamp
SB_MESSAGE.M_START_TIME - date
SELECT SB_BATCH.B_UID, SB_BATCH.B_CREATE_DT, SB_MESSAGE.M_START_TIME,
to_date(to_char(SB_BATCH.B_CREATE_DT), 'DD-MON-RR HH24:MI:SS') as time_in_minutes
FROM SB_BATCH, SB_MESSAGE
WHERE
SB_BATCH.B_UID = SB_MESSAGE.M_B_UID;
Result:
Error report -
SQL Error: ORA-01830: date format picture ends before converting entire input string
01830. 00000 - "date format picture ends before converting entire input string"
You can subtract two timestamps to get an INTERVAL DAY TO SECOND, from which you calculate how many minutes elapsed between the two timestamps. In order to convert SB_MESSAGE.M_START_TIME to a timestamp you can use CAST.
Note that I have also removed your implicit table join with an explicit INNER JOIN, moving the join condition to the ON clause.
SELECT t.B_UID,
t.B_CREATE_DT,
t.M_START_TIME,
EXTRACT(DAY FROM t.diff)*24*60 +
EXTRACT(HOUR FROM t.diff)*60 +
EXTRACT(MINUTE FROM t.diff) +
ROUND(EXTRACT(SECOND FROM t.diff) / 60.0) AS diff_in_minutes
FROM
(
SELECT SB_BATCH.B_UID,
SB_BATCH.B_CREATE_DT,
SB_MESSAGE.M_START_TIME,
SB_BATCH.B_CREATE_DT - CAST(SB_MESSAGE.M_START_TIME AS TIMESTAMP) AS diff
FROM SB_BATCH
INNER JOIN SB_MESSAGE
ON SB_BATCH.B_UID = SB_MESSAGE.M_B_UID
) t
Convert the timestamp to a date using cast(... as date). Then take the difference between the dates, which is a number - expressed in days, so if you want it in minutes, multiply by 24*60. Then round the result as needed. I made up a small example below to isolate just the steps needed to answer your question. (Note that your query has many other problems, for example you didn't actually take a difference of anything anywhere. If you need help with your query in general, please post it as a separate question.)
select ts, dt, round( (sysdate - cast(ts as date))*24*60, 2) as time_diff_in_minutes
from (select to_timestamp('2016-08-23 03:22:44.734000', 'yyyy-mm-dd hh24:mi:ss.ff') as ts,
sysdate as dt from dual )
;
TS DT TIME_DIFF_IN_MINUTES
-------------------------------- ------------------- --------------------
2016-08-23 03:22:44.734000000 2016-08-23 08:09:15 286.52
I am trying to reconfigure a View into a materialized view for summary roll-ups. IS there a way to implement the process so that I can gather each months parameters from the materialized view easily without having to run Substitute variables each time:
def RECORDMONTH = 201402
def BeginDate = 2/01/2014
def EndDate = 2/28/2014
def YEAR_MONTH = 2014-02
Would it be able to be designed in a way that can pull all twqeleve months even if that month has happened yet.
Really appreciate it.
I'm guessing that this is a continuation of a number of other questions that you've asked. My guess is that you just want something like this where you generate the months you want (in this case the 12 months of 2015) and use that in an outer join in your query.
WITH all_months AS
SELECT add_months( date '2015-01-01', level-1 ) start_of_month,
last_day( add_months( date '2015-01-01', level-1 )) end_of_month
FROM dual
CONNECT BY level <= 12
)
SELECT am.start_of_month,
count(*) num_rows,
<<other aggregates>>
FROM your_table yt
RIGHT OUTER JOIN all_months am
ON( yt.some_date BETWEEN am.start_of_month AND am.end_of_month )
GROUP BY am.start_of_month
I want to Group the rows based on certain columns, i.e. if data is same in these columns in continuous rows, then assign same Group Number to them, and if its changed, assign new one. This become complex as the same data in the columns could appear later in some other rows, so they have to be given another Group Number as they are not in continuous rows with previous group.
I used cte for this purpose and it is giving correct output also, but is so slow that iterating over 75k+ rows takes about 15 minutes. The code I used is:
WITH
cte AS (SELECT ROW_NUMBER () OVER (ORDER BY Patient_ID, Opnamenummer, SPECIALISMEN, Opnametype, OntslagDatumTijd) AS RowNumber,
Opnamenummer, Patient_ID, AfdelingsCode, Opnamedatum, Opnamedatumtijd, Ontslagdatum, Ontslagdatumtijd, IsSpoedopname, OpnameType, IsNuOpgenomen, SpecialismeCode, Specialismen
FROM t_opnames)
SELECT * INTO #ttt FROM cte;
WITH cte2 AS (SELECT TOP 1 RowNumber,
1 AS GroupNumber,
Opnamenummer, Patient_ID, AfdelingsCode, Opnamedatum, Opnamedatumtijd, Ontslagdatum, Ontslagdatumtijd, IsSpoedopname, OpnameType, IsNuOpgenomen, SpecialismeCode, Specialismen
FROM #ttt
ORDER BY RowNumber
UNION ALL
SELECT c1.RowNumber,
CASE
WHEN c2.Afdelingscode <> c1.Afdelingscode
OR c2.Patient_ID <> c1.Patient_ID
OR c2.Opnametype <> c1.Opnametype
THEN c2.GroupNumber + 1
ELSE c2.GroupNumber
END AS GroupNumber,
c1.Opnamenummer,c1.Patient_ID,c1.AfdelingsCode,c1.Opnamedatum,c1.Opnamedatumtijd,c1.Ontslagdatum,c1.Ontslagdatumtijd,c1.IsSpoedopname,c1.OpnameType,c1.IsNuOpgenomen, SpecialismeCode, Specialismen
FROM cte2 c2
JOIN #ttt c1 ON c1.RowNumber = c2.RowNumber + 1
)
SELECT *
FROM cte2
OPTION (MAXRECURSION 0) ;
DROP TABLE #ttt
I tried to improve performance by putting output of cte in a temp table. That increased the performance, but still its too slow. So, how can I increase the performance of this code to run it under 10 seconds for 75k+ records? The output before cancelling the query is: Screenshot. As visible from the image, data is same in columns Afdelingscode,Patient_ID and Opnametype in RowNumber 3,5 and 6, but they have different GroupNumber because of concurrency of the rows.
Without data its not that easy to test but i would try first to not use temporary table and just use both cte from start to end, ie;
;WITH
cte AS (...),
cte2 AS (...)
select * from cte2
OPTION (MAXRECURSION 0);
Without knowing indices etc... for instance, you do a lot of ordering in the first cte. Is this supported by indices (or one multicolumn index) or not?
Without the data i don't have the option to play with it but looking at this:
CASE
WHEN c2.Afdelingscode <> c1.Afdelingscode
OR c2.Patient_ID <> c1.Patient_ID
OR c2.Opnametype <> c1.Opnametype
THEN c2.GroupNumber + 1
ELSE c2.GroupNumber
i would try to take a look at partition by statement in row_number
So try to run this:
WITH
cte AS (
SELECT ROW_NUMBER () OVER (PARTITION BY Afdelingscode , Patient_ID ,Opnametype ORDER BY Patient_ID, Opnamenummer, SPECIALISMEN, Opnametype, OntslagDatumTijd ) AS RowNumber,
Opnamenummer, Patient_ID, AfdelingsCode, Opnamedatum, Opnamedatumtijd, Ontslagdatum, Ontslagdatumtijd, IsSpoedopname, OpnameType, IsNuOpgenomen
FROM t_opnames)
I have a query which I need to make more efficient.
I am breaking it down into sections to see where the efficiency floors are, I currently have a few Nested Select statements, are these a performance problem?
Here is an example of one of them:
SELECT AgreementID,
DueDate,
UpdatedAmountDue AS AmountDue,
COALESCE((SELECT SUM(UpdatedAmountDue)
FROM RepaymentBreakdown AS B
WHERE CONVERT(datetime, CONVERT(varchar, DueDate, 103), 103) <=
CONVERT(datetime, CONVERT(varchar, R.DueDate, 103), 103)
AND B.AgreementID = R.AgreementID),0) AS DueTD,
RN=ROW_NUMBER() OVER (Partition BY R.AgreementID ORDER BY DueDate)
FROM RepaymentBreakdown AS R
Is there a more clean and efficient way of getting the data of DueTD?
Basically, for each line of a repayment schedule result, I want to get:
AgreementID,
DueDate,
AmountDue,
AmountDueToDate (DueTD)
RowNumber.
The table I am querying is structured as follows:
AgreementID (int),
DueDate (datetime),
AmountDue (decimal(9,2)),
UpdatedAmountDue (decimal(9,2))*
*UpdatedAmountDue is always referenced as it is the moving figure, AmountDue is always fixed, as a reference value.
So, I think you could get performance boost just by removing convert, like this:
select
AgreementID,
DueDate,
UpdatedAmountDue as AmountDue,
(
select sum(B.UpdatedAmountDue)
from RepaymentBreakdown as B
where B.DueDate <= R.DueDate and B.AgreementID = R.AgreementID
) as UpdatedAmountDue
from RepaymentBreakdown AS R
The fastest way I know to calculate running total in SQL Server 2008 would be to use recursive CTE, see my answer here Calculate a Running Total in SqlServer. In your case the query would be smth like this:
create table #t (....., primary key (AgreementID, ord))
insert into #t (AgreementID, DueDate, UpdatedAmountDue, ord)
select AgreementID, DueDate, UpdatedAmountDue, row_number() over (partition by AgreementID, DueDate order by DueDate asc)
;with
CTE_RunningTotal
as
(
select T.ord, T.AgreementID, T.DueDate, T.UpdatedAmountDue as T.AmountDue, T.UpdatedAmountDue
from #t as T
where T.ord = 1
union all
select T.ord, T.AgreementID, T.DueDate, T.UpdatedAmountDue as T.AmountDue, T.UpdatedAmountDue + C.UpdatedAmountDue as UpdatedAmountDue
from CTE_RunningTotal as C
inner join #t as T on T.ord = C.ord + 1 and T.AgreementID = C.AgreementID
)
select AgreementID, DueDate, AmountDue, UpdatedAmountDue
from CTE_RunningTotal as C
option (maxrecursion 0)
Your conversion of the datetime to a date has several issues.
First, it is not guaranteed to always produce correct results depending on your servers language settings. If you need to do String manipulation on a datetime value always use CONVERT(,,126).
But more importantly, it prevents index usage. Instead use CAST(DueDate AS DATE) as the optimizer recognizes that conversion to be index-safe.
Afterwards you might want to add an index on AgreementId,DueDate and either INCLUDE UpdatedAmountDue or better make it clustered.
Assuming UpdatedAmountDue cannot be NULL, you can get rid of the COALESCE too, as the sum always includes the current row.