I have a Device that send event on a SQL data base. The event a log like this:
index|eventNumber|TriggerTime|RecoverTime
The issue I have is if they have a power off/on on the device it create sont unrecover event like this
index|eventNumber|TriggerTime|RecoverTime
1|2|2020-04-03 8:00|2020-04-03 8:10
2|3|2020-04-03 8:00|2020-04-03 8:30
3|3|2020-04-03 8:20|NULL
4|2|2020-04-03 8:21|2020-04-03 8:23
5|3|2020-04-03 8:40|NULL
What I want to do is to copy the triggerTime of the next event with the same eventNumber in the recoverTime. Like this
index|eventNumber|TriggerTime|RecoverTime
1|2|2020-04-03 8:00|2020-04-03 8:10
2|3|2020-04-03 8:00|2020-04-03 8:30
3|3|2020-04-03 8:20|2020-04-03 8:40
4|2|2020-04-03 8:21|2020-04-03 8:23
5|3|2020-04-03 8:40|NULL
What is the best way to do this?
You can use lead():
select el.*,
coalesce(recovertime,
lead(triggertime) over (partition by eventnumber order by triggertime)
) as new_recovertime
from eventlog;
If you need this in an update, you can use join
update eventlog el join
(select el.*,
lead(triggertime) over (partition by eventnumber order by triggertime) as new_recovertime
from eventlog
) el2
on el2.eventnumber = el.eventnumber and
el2.eventtime = el.eventtime
set recovertime = new_recovertime
where recovertime is null and new_recovertime is not null;
the mysql server is 5.7, so it doesn't have the lead fonction. I find a way to do a select to that add a colone with the value to put in the null place, but I don't know how to put it in an update.
(SELECT
event_index, event_log_index, `trigger_time#timestamp`, `recover_time#timestamp`
, (SELECT
`trigger_time#timestamp`
FROM
plc.`test` alfabetInner
WHERE
alfabetInner.`trigger_time#timestamp` > plc.`test`.`trigger_time#timestamp` AND alfabetInner.`event_log_index` = plc.`test`.`event_log_index`
ORDER BY
`trigger_time#timestamp` ASC
LIMIT 0, 1
) AS 'Nexttrigger_time#timestamp'
FROM
plc.`test`
ORDER BY
`trigger_time#timestamp` ASC)
Related
I have read a bunch of ways that has gotten me this far. But I can't get to the finish line.
I have a table of coupon codes. I want to use one transaction to select the next available code, mark it as used and input the order number. I can get the update and nested select to work, but I cannot figure out how to actually return the coupon code from the select. It just returns 1 row updated.
Here's what I've got:
UPDATE `prcoupon` pr
SET
`pr`.`status` = '1',
`pr`.`invoicenumber` = '09990002'
WHERE
`pr`.`couponCode` = (SELECT
`prcoupon`.`couponcode`
FROM
`prcoupon`
WHERE
`status` = 0
LIMIT 1)
Sample data
What I need returned is: couponCode: SL2T-03A0-JVCY-W2XMXG
If I understand correctly, you can try to use UPDATE ... JOIN with ROW_Nunber windwon function.
UPDATE prcoupon pr
JOIN (
SELECT *,ROW_NUMBER() OVER(ORDER BY couponCode) rn
FROM prcoupon
WHERE status = 0
) t2 ON pr.couponcode = t2.couponcode
SET pr.status = 1,
pr.invoicenumber = '09990002'
WHERE rn = 1
sqlfiddle
I have a complicated SQL query running on a MSSQL server. I need to change it to make it work also on a MySQL database, however I do not know how to do this. The query is the following and is used to build a complicated Sankey graph in a BI tool
SELECT Id, Delivery_Date, Queue, Step1, Step2, Step3, Step4, Step5, Step6, Step7,
CONVERT(DECIMAL(16,16),RAND((CHECKSUM(NEWID())))*(1 - 0.999999999)+0.999999999) as Counter
FROM
(SELECT t.ID as Id, t.DELIVERY_DATE as Delivery_Date, t.QUEUE as Queue, CONCAT(t.EVENT_NAME,' ', t.EVENT_STATUS, ' ',t .ENDING) as edsda,
'Step' + cast(row_number() over (partition by ID order by ID)
as varchar(10)) ColumnSeq
FROM (SELECT DISTINCT e.[ID] AS ID,
e.[DIALOG_DELIVERY_DATE] AS DELIVERY_DATE,
e.[QUEUE_ID] as QUEUE,
s2.[STEP_NUMBER] AS STEP_NUMBER,
s2.[CURRENT_EVENT] AS EVENT_NAME,
s2.[EVENT_STATUS] as EVENT_STATUS,
e.[DELIVERY_END] AS DELIVERY_END,
CASE WHEN s2.[STEP_NUMBER] = max(s2.[STEP_NUMBER]) over (partition by s2.[ID])
and s2.[EVENT_NUMBER] = min(s2.[EVENT_NUMBER]) over (partition by s2.[ID], s2.[STEP_NUMBER]) THEN e.[DELIVERY_END] ELSE NULL END as ENDING
FROM [dbo].[DELIVERY_STEP_EVENTS] s1,
[dbo].[DELIVERY_STEP_EVENTS] s2,
[dbo].[DELIVERY_EVENTS] e
WHERE s1.[ID] = s2.[ID]
AND e.[ID] = s2.[ID]
AND s1.[EVENT_NAME] = s2.[EVENT_NAME]
AND s1.[STEP_NUMBER] <= s2.[STEP_NUMBER] AS t) AS temp
pivot
(max(edsda)
for ColumnSeq in (Step1,Step2,Step3,Step4, Step5, Step6, Step7)
)
AS Piv
My biggest issue so far has been the RAND((CHECKSUM(NEWID()))) which I do not know how to change to MySQL.
Any help would be greatly appreciated.
My biggest issue so far has been the RAND((CHECKSUM(NEWID()))) which I
do not know how to change to MySQL.
RAND((CHECKSUM(NEWID()))) should behave identical to RAND() function in MySQL. So you could simply replace this:
CONVERT(DECIMAL(16,16),RAND((CHECKSUM(NEWID())))*(1 - 0.999999999)+0.999999999)
With this:
CAST(RAND() * (1 - 0.999999999) + 0.999999999 AS DECIMAL(16, 16))
SQL Server
MySQL
I'm looking for a way to order my results based on the actual time. In my table yo can see values like:
1,23:45
2,9:45
3,27:43
When I do a query I would like to know how to order them based on their actual 24 hour time.
Ex:
3,3:43
2,9:45
1,23:45
Notice how it changes 27:43 to 3:43, and creates the order.
Where I am using it, in this query:
SELECT *,COALESCE(ADDTIME(s.`departure_time`,SEC_TO_TIME(rt.delay)),s.`departure_time`) as `rt_time` FROM `stop_times` s INNER JOIN `trips` t ON s.`trip_id` = t.`trip_id` INNER JOIN `stops` st ON st.`stop_id` = s.`stop_id` INNER JOIN `routes` r ON r.`route_id` = t.`route_id` LEFT JOIN `rt_trips` rt ON t.`trip_id` = rt.`trip_id` where (s.`stop_id` = 'CB900') and ( ( s.`departure_time` >= '00:50' and s.`departure_time` <= '05:50') OR ( s.`departure_time` >= '24:50' and s.`departure_time` <= '29:50') ) and (s.`pickup_type` = '0') and (t.`service_id` IN ('removed to make it easier')) HAVING (`rt_time` BETWEEN '01:50' and '05:50' ) ) OR ( `rt_time` BETWEEN '25:50' and '29:50' ) ORDER BY `order` ASC
Explanation:
Information is a transit schedule, that may go forward onto the next day which may be a saturday. So, times may become 25:50, where that means 1:50 the next day.
Thanks
Cyrus
Hmmm, if you just want to get a value between 0 and 24 hours, then I would do:
select concat(mod(substring_index(time_column, ':', 1) + 0, 24), ':',
substring_index(time_column, ':', -1)
)
Try this function on the time_column
concat(mod(substr(time_column,1,INSTR(time_column, ':')-1),24)
,substr(time_column,INSTR(time_column, ':'),3)
)
You might need to cast date to string to integer, do the maths, and again cast it to time. But the fiddle version seems to work properly on varchar to integer conversion. Check this
http://sqlfiddle.com/#!9/ff60f9/1
I have multiple update statements in a stored procedure (as shown below).
Question is I am trying to combine them into one UPDATE statement as there is a performance issue (takes longer to execute stored procedure). I tried putting columns (such as PONUMBER, VENDORID etc) in a single update statement but it is throwing errors.
Please suggest.
UPDATE rptMaster SET PONUMBER = (select top 1 poMaster.PONUMBER from poMaster where poMaster.ITEMNMBR =rptMaster.ITEMNMBR and
poMaster.UnCommited > 0)
UPDATE rptMaster SET VENDORID = (select top 1 poMaster.VENDORID from poMaster where poMaster.ITEMNMBR =rptMaster.ITEMNMBR and
poMaster.UnCommited > 0)
UPDATE rptMaster SET DUEDATE = (select top 1 poMaster.REQDATE from poMaster where poMaster.ITEMNMBR =rptMaster.ITEMNMBR and
poMaster.UnCommited > 0)
UPDATE rptMaster SET POQTYORDER = (select top 1 (poMaster.QTYORDER / rptMaster.UOMQTY) from poMaster where poMaster.ITEMNMBR =rptMaster.ITEMNMBR and
poMaster.UnCommited > 0)
Mine is similar to polkduran's:
WITH PO AS (
SELECT PONUMBER
, VENDORID
, REQDATE
, QTYORDER
, ITEMNMBR
, ROW_NUMBER() OVER (PARTITION BY ITEMNMBR ORDER BY ??) as RN
FROM poMaster
WHERE UnCommited > 0
)
UPDATE rptMaster
SET PONUMBER = po.PONUMBER
, VENDORID = po.VENDORID
, DUEDATE = po.REQDATE
, POQTYORDER = po.QTYORDER / rptMaster.UOMQTY
FROM rptMaster
JOIN PO
ON PO.ITEMNMBR = rptMaster.ITEMNMBR
and PO.RN = 1
I'm using a Common Table Expression (CTE) to assign a row number to each poMaster record, with the records for each value of ITEMNMBR numbered separately. This allows us to pick to the first record for each ITEMNBR in our JOIN, later, similar to the way you were using Top 1 in your subqueries.
Please note, though: because you didn't indicate how you wanted to select the Top 1 record in your query, I had to leave the ORDER BY clause in the CTE unspecified. (I put ?? in as a placeholder.) You need to specify one or more sort fields in place of the ?? so it knows how to sort and number the records.
You can make an update using a join clause:
update rpt
set
PONUMBER = po.PONUMBER,
VENDORID = po.VENDORID,
DUEDATE = po.REQDATE,
POQTYORDER = (po.QTYORDER / rpt.UOMQTY)
from rptMaster rpt
inner join poMaster po
on po.ITEMNMBR = rpt.ITEMNMBR
where po.UnCommited > 0
I don't have a way to test it right now but that might work.
I've created this temporal table in my store procedure, as you can see I have more than 1 records for the same ID:
#tmpTableResults
TmpInstallerID TmpConfirmDate TmpConfirmLocalTime
============== ============== ===================
173 2011-11-08 11:45:50
278 2011-11-04 09:06:26
321 2011-11-08 13:21:35
321 2011-11-08 11:44:54
483 2011-11-08 11:32:00
483 2011-11-08 11:31:59
645 2011-11-04 10:03:15
645 2011-11-04 07:03:15
That is the result of the query to create #tmpTableResults
DECLARE #tmpTableResults TABLE
(
TmpInstallerID int,
TmpConfirmDate date,
TmpConfirmLocalTime time
)
DECLARE #tmpTableQuery VarChar(800)
SET #tmpTableQuery = 'select FxWorkorder.INSTALLERSYSID, FxWorkorder.CONFIRMDATE, FxWorkorder.CONFIRMLOCALTIME from FxWorkorder
join install on FxWorkorder.INSTALLERSYSID = install.sysid
join RouteGroupWorkarea on FxWorkorder.WORKAREAGROUPSYSID = RouteGroupWorkarea.IWORKAREA_ID
join RoutingGroup on RouteGroupWorkarea.IRG_ID = RoutingGroup.IRG_IDENTITY
where FxWorkorder.SCHEDULEDDATE > = #StartDate and FxWorkorder.SCHEDULEDDATE <= #EndDate
and FxWorkorder.Jobstatus <> "Unassign"
and FxWorkorder.Jobstatus <> "Route"
and install.FOXTELCODE <> ""
and FxWorkorder.CONFIRMLOCALTIME is not null
and FxWorkorder.CONFIRMDATE <> ""
group by FxWorkorder.INSTALLERSYSID, FxWorkorder.CONFIRMDATE, FxWorkorder.CONFIRMLOCALTIME
order by FxWorkorder.INSTALLERSYSID, FxWorkorder.CONFIRMDATE, FxWorkorder.CONFIRMLOCALTIME desc '
INSERT INTO #tmpTableResults EXEC(#tmpTableQuery)
I'm creating another query to get data from another table and only the first record from the temporal table for the same INSTALLERSYSID
SELECT RoutingGroup.SDESCRIPTION, FxWorkorder.INSTALLERSYSID, FxWorkOrder.JOBSTATUS, Install.FOXTELCODE,
install.NAME, FxWorkOrder.ScheduledDate,
count(*) as TotalJobs, COUNT(CONFIRMDATE) as ConfirmedJobs,
(select TmpInstallerID, TmpConfirmDate, TmpConfirmLocalTime from #tmpTableResults where TmpInstallerID = FxWorkorder.INSTALLERSYSID)
from FxWorkorder
join install on fxworkorder.INSTALLERSYSID = install.sysid
join RouteGroupWorkarea on FxWorkOrder.WORKAREAGROUPSYSID = RouteGroupWorkarea.IWORKAREA_ID
join RoutingGroup on RouteGroupWorkarea.IRG_ID = RoutingGroup.IRG_IDENTITY
where FxWorkorder.SCHEDULEDDATE > = #StartDate and FxWorkorder.SCHEDULEDDATE <= #EndDate
and FxWorkOrder.Jobstatus <> 'Unassign'
and FxWorkOrder.Jobstatus <> 'Route'
and Install.FOXTELCODE <> ''
group by RoutingGroup.SDESCRIPTION,FxWorkOrder.INSTALLERSYSID, FxWorkOrder.JOBSTATUS, Install.FOXTELCODE,install.NAME, FxWorkOrder.ScheduledDate,FxWorkOrder.WORKAREAGROUPSYSID
When I tried to save the sp I got the error
"Only one expression can be specified in the select list when the subquery is not introduced with EXISTS."
I can't see why I got this error. But if I run the query in sql that works. Can someone see the error?
I don't know how your second query works for you ‘in sql’ (where is that supposed to be? do you mean SSMS = SQL Server Management Studio?), but I'm sure it cannot possibly work in any version of SQL Server that exists at the moment. It's because of this subquery in the SELECT list:
(select TmpInstallerID, TmpConfirmDate, TmpConfirmLocalTime from #tmpTableResults where TmpInstallerID = FxWorkorder.INSTALLERSYSID)
The thing is, every expression in the SELECT clause should be scalar, but this subquery returns a row of more than one value. Even if it's only one row, it is illegal there, because it returns several columns. The subquery in that context should return no more than one value, i.e. it should be one column and the result produced should contain either no rows or just one.
You could try this query instead (although I'm not entirely sure without knowing more details about your schema):
SELECT
RoutingGroup.SDESCRIPTION,
FxWorkorder.INSTALLERSYSID,
FxWorkOrder.JOBSTATUS,
Install.FOXTELCODE,
install.NAME, FxWorkOrder.ScheduledDate,
count(*) as TotalJobs, COUNT(CONFIRMDATE) as ConfirmedJobs,
tmp.TmpInstallerID,
tmp.TmpConfirmDate,
tmp.TmpConfirmLocalTime
from FxWorkorder
join install on fxworkorder.INSTALLERSYSID = install.sysid
join RouteGroupWorkarea on FxWorkOrder.WORKAREAGROUPSYSID = RouteGroupWorkarea.IWORKAREA_ID
join RoutingGroup on RouteGroupWorkarea.IRG_ID = RoutingGroup.IRG_IDENTITY
join #tmpTableResults tmp ON tmp.TmpInstallerID = FxWorkorder.INSTALLERSYSID
where FxWorkorder.SCHEDULEDDATE > = #StartDate
and FxWorkorder.SCHEDULEDDATE <= #EndDate
and FxWorkOrder.Jobstatus <> 'Unassign'
and FxWorkOrder.Jobstatus <> 'Route'
and Install.FOXTELCODE <> ''
group by
RoutingGroup.SDESCRIPTION,
FxWorkOrder.INSTALLERSYSID,
FxWorkOrder.JOBSTATUS,
Install.FOXTELCODE,install.NAME,
FxWorkOrder.ScheduledDate,
FxWorkOrder.WORKAREAGROUPSYSID
tmp.TmpInstallerID,
tmp.TmpConfirmDate,
tmp.TmpConfirmLocalTime
That is, I added one more join, the one to #tmpTableResults, as well as added the columns you were trying to pull to the SELECT clause and to the GROUP BY clause.
Also, if I were you I would consider using short aliases for tables, like this:
SELECT
…
wo.INSTALLERSYSID,
wo.JOBSTATUS,
…
from FxWorkorder wo
join …
That might make your queries more readable.