I know there are a lot of other SO entries that seem like this one, but I haven't found one that actually answers my question so hopefully one of you can either answer it or point me to another SO question that is related.
Basically, I have the following query that returns Venues that have any CheckIns that contain the searched Keyword ("foobar" in this example).
SELECT DISTINCT v.*
FROM "venues" v
INNER JOIN "check_ins" c ON c."venue_id" = v."id"
INNER JOIN "keywordings" ks ON ks."check_in_id" = c."id"
INNER JOIN "keywords" k ON ks."keyword_id" = k."id"
WHERE (k."name" = 'foobar')
I want to SELECT and ORDER BY the count of the matched Keyword for each given Venue. E.g. if there have been 5 CheckIns that have been created, associated with that Keyword, then there should be a returned column (called something like keyword_count) with the value 5 which is sorted.
Ideally this should be done without any queries in the SELECT clause, or preferably none at all.
I've been struggling with this for a while and my mind is just going blank (perhaps it's been too long a day) so some help would be greatly appreciated here.
Thanks in advance!
Sounds like you need something like:
SELECT v.x, v.y, count(*) AS keyword_count
FROM "venues" v
INNER JOIN "check_ins" c ON c."venue_id" = v."id"
INNER JOIN "keywordings" ks ON ks."check_in_id" = c."id"
INNER JOIN "keywords" k ON ks."keyword_id" = k."id"
WHERE (k."name" = 'foobar')
GROUP BY v.x, v.y
ORDER BY 3
Related
SELECT
A.PACKINGSLIPID, A.ITEMID, A.QTY,
D.INVENTBATCHID, C.ORIGPURCHID, c.QTY, C.PURCHUNIT
FROM
INVENTTRANS A
LEFT JOIN
VENDPACKINGSLIPTRANS C ON A.PACKINGSLIPID = C.PACKINGSLIPID
INNER JOIN
INVENTDIM D ON A.INVENTDIMID = D.INVENTDIMID
WHERE
C.ORIGPURCHID = 'PO-SIZA-000003809'
AND A.DATAAREAID = C.DATAAREAID
GROUP BY
A.PACKINGSLIPID, A.ITEMID, A.QTY, D.INVENTBATCHID,
C.ORIGPURCHID, c.qty, C.PURCHUNIT
This looks like Dynamics AX. Your duplicates are coming from the fact that VendPackingSlipTrans and InventTrans is a many to many relationship.
As commenters have pointed out, specifying in the where clause that the left joined table contain a value is the same as making it an inner join. My guess is that they thing you're really trying to do is this:
SELECT
VENDPACKINGSLIPTRANS.PACKINGSLIPID,
VENDPACKINGSLIPTRANS.ITEMID,
VENDPACKINGSLIPTRANS.QTY,
INVENTDIM.INVENTBATCHID,
VENDPACKINGSLIPTRANS.ORIGPURCHID,
VENDPACKINGSLIPTRANS.PURCHUNIT
FROM VENDPACKINGSLIPTRANS
INNER JOIN INVENTDIM
ON INVENTDIM.[PARTITION] = VENDPACKINGSLIPTRANS.[PARTITION]
AND INVENTDIM.DATAAREAID = VENDPACKINGSLIPTRANS.DATAAREAID
AND INVENTDIM.INVENTDIMID = VENDPACKINGSLIPTRANS.INVENTDIMID
WHERE VENDPACKINGSLIPTRANS.ORIGPURCHID = 'PO-SIZA-000003809'
Notice I'm joining on Partition, DataAreaId, and InventDimId. That's because all the AX indices except the RECID index have partition at the beginning. Since InventDim has a unique index on Partition/DataAreaId/InventDimId, I know I'm not going to get any duplicates, every line is vendor packing slip line.
If you neeeeed the InventTrans, the VendPackingSlipTrans does have an InventTransId field, but remember there is not necessarily one InventTrans record per InventTransId. You can't obviate the need to understand the data model by chucking a bunch of group bys at the problem.
long time lurker. I've searched and searched though none of the solutions work for me.
I'm working in a Sybase (ASE) db (most mssql/mysql transactional db solutions will work just fine)
In my example, I'm trying to calculate/count the number of times a specific 'party_id' is listed in a column. The problem I'm having is that it's only counting FOR each row- so of course the count is always going to be 1.
See output:
(I would like for party_id 130568 to show '2' in the refs column, 125555 to show '5', etc.)
output
Here is my query:
select
count(distinct p.party_id) as refs,
p.party_id,
sp_first_party(casenum),
c.casenum,
mld.mailing_list,
p.our_client
from cases c
inner join party p on c.casenum=p.case_id
inner join names n on n.names_id=p.party_id
inner join mailing_list_defined mld on n.names_id=mld.names_id
where
mld.mailing_list like 'Mattar Stars'
and mld.addr_type like 'Home'
and n.deceased='N'
and p.our_client='Y'
group by p.party_id, c.casenum, mld.mailing_list, p.our_client
order by sp_first_party(casenum) asc
Any tips would be greatly appreciated.
Thank you
Sounds like you need to be using an APPLY statement. Not sure if the join criteria on the APPLY statement is correct, but you should be able to extrapolate the logic. See if that will work with Sybase.
SELECT pic.PartyInstanceCount AS refs
,p.party_id
,sp_first_party(casenum)
,c.casenum
,mld.mailing_list
,p.our_client
FROM cases AS c
INNER JOIN party AS p ON c.casenum = p.case_id
INNER JOIN names AS n ON n.names_id = p.party_id
INNER JOIN mailing_list_defined AS mld ON n.names_id = mld.names_id
OUTER APPLY (
SELECT COUNT(1) AS PartyInstanceCount
FROM party p2
WHERE p2.case_id = c.casenum
) pic
WHERE mld.mailing_list LIKE 'Mattar Stars'
AND mld.addr_type LIKE 'Home'
AND n.deceased = 'N'
AND p.our_client = 'Y'
ORDER BY
sp_first_party(casenum) ASC
I made a query and wanted to not have any duplicates but i got some times 3 duplicates and when i used DISTINCT or DISTINCTROW i got only 2 duplicates.
SELECT f.flight_code,
f.status,
a.airport_name,
a1.airport_name,
f.departing_date+f.departing_time AS SupposedDepartingTime,
f.landing_date+f.landing_time AS SupposedLandingTime,
de.actual_takeoff_date+de.actual_takeoff_time AS ActualDepartingTime,
SupposedLandingTime+(ActualDepartingTime-SupposedDepartingTime) AS ActualLandingTime
FROM
(((Flights AS f
LEFT JOIN Aireports AS a
ON a.airport_code = f.depart_ap)
LEFT JOIN Aireports AS a1
ON f.target_ap = a1.airport_code)
LEFT JOIN Irregular_Events AS ie
ON f.flight_code = ie.flight_code)
LEFT JOIN Delay_Event AS de
ON ie.IE_code = de.delay_code;
had to use LEFT JOIN because when i used INNER JOIN i missed some of the things i wanted to show because i wanted to see all the flights and not only the flights that got delayed or canceled.
This is the results when i used INNER JOIN, you can see only the flights that have the status "ביטול" or "עיכוב" and that is not what i wanted.
[the results with LEFT JOIN][2]
[2]: https://i.stack.imgur.com/cgE2G.png
and when i used DISTINCT where you see the rows with the NUMBER 6 on the first column it appear only two times
IMPORTANT!
I just checked my query and all the tables i use there and i saw my problem but dont know how to fix it!
in the table Irregular_Events i have more the one event for flights 3,6 and 8 and that is why when i use LEFT JOIN i see more even thou i use distinct, please give me some help!
Not entirely sure without seeing the table structure, but this might work:
SELECT f.flight_code,
f.status,
a.airport_name,
a1.airport_name,
f.departing_date+f.departing_time AS SupposedDepartingTime,
f.landing_date+f.landing_time AS SupposedLandingTime,
de.actual_takeoff_date+de.actual_takeoff_time AS ActualDepartingTime,
SupposedLandingTime+(ActualDepartingTime-SupposedDepartingTime) AS ActualLandingTime
FROM
((Flights AS f
LEFT JOIN Aireports AS a
ON a.airport_code = f.depart_ap)
LEFT JOIN Aireports AS a1
ON f.target_ap = a1.airport_code)
LEFT JOIN
(
SELECT
ie.flight_code,
de1.actual_takeoff_date,
de1.actual_takeoff_time
FROM
Irregular_Events ie
INNER JOIN Event AS de1
ON ie.IE_code = de1.delay_code
) AS de
ON f.flight_code = de.flight_code
It is hard to tell what is the problem with your query without any sample of the output, and without any description of the structure of your tables.
But your problem is that your are querying from the flights table, which [I assume] can be linked to multiple irregular_events, which can possibly also be linked to multiple delay_event.
If you want to get only one row per flight, you need to make sure your joins return only one row too. Maybe you can do it by adding one more condition to the join, or by adding a condition in a sub-query.
EDIT
You could try to add a GROUP BY to the query:
GROUP BY
f.flight_code,
f.status,
a.airport_name,
a1.airport_name;
I am trying to make a summary table of all the items I have. I have a raw data table with 10 users who respectively have different items. There are maximum 3 different items and I want to do a count to see how many items each individual has. The following is my code.
Select b.Country,b.UserID,Num_including_fruits, Apple,Orange
from
(((SELECT o.Country,o.UserID, IIF(ISNULL(Count(o.UserID)),0,Count(o.UserID))
AS Num_including_fruits
FROM [SEA2_View] as o
GROUP BY o.UserID, o.Country
ORDER BY Country)as b
LEFT JOIN
(SELECT o.Country,o.UserID,IIF(ISNULL(Count(o.UserID)),0,Count(o.UserID)
AS Apple
FROM [APAC2_View] as o
WHERE o.fruit_status <>"fresh" AND o.HWType = "Apple"
GROUP BY o.Country,o.UserID)as d
ON (b.UserID = d.UserID))
LEFT JOIN
(SELECT o.Country,o.UserID,IIF(ISNULL(Count(o.UserID)),0,Count(o.UserID))
AS Orange
FROM [SEA2_View] as o
WHERE o.fruit_status <>"fresh" AND o.HWType = "Orange"
GROUP BY o.Country,o.UserID)as e
ON (d.UserID = e.UserID))
;
The first join returns the correct result but the second join somehow returns all 0, which is incorrect. Therefore please help! and I would appreciate any advice for best practice when it comes to joins in SQL. Thanks lot!
Are you sure you don't have a table naming error?
You're first joining [SEA2_View] with [APAC2_View]. The second join is joining with [SEA2_View] with itself.
I have this code:
select Users.phoneMac, Users.apMac, Locations.Lon, Locations.Lat
from Locations, Users
inner join (
select u.phoneMac, max(u.strenght) as most
from Users u, Locations l
where u.apMac = l.apMac
group by u.phoneMac
) as ij on ij.phoneMac=Users.phoneMac and Users.strenght = ij.most
where Locations.apMac = Users.apMac;
It worked for me fine but when I added more data to users table this query calculated results from all the data and I wanted to get results just from latest data. So I added timestamp to Users table.
So can you help me fix this code so it first take only data from latest timestamp for every user(users.phoneMac)(there can be more then 1 row of data for same phoneMac) and then do the rest of calculations.
You're already picking the max value of the "strenght" field and joining on that, so why not use the same approach again for your timestamp field? Something like:
SELECT Users.phoneMac, Users.apMac, Locations.Lon, Locations.Lat
FROM Locations
INNER JOIN Users
ON Users.apMac = Locations.apMac
INNER JOIN (
SELECT u.phoneMac, max(u.strenght) AS most
FROM Locations l
INNER JOIN Users u ON u.apMac = l.apMac
GROUP BY u.phoneMac) AS ij
ON ij.phoneMac = Users.phoneMac
AND Users.strenght = ij.most
INNER JOIN (
SELECT u2.phoneMac, max(u2.timestampfield) AS latest
FROM Locations l2
INNER JOIN Users u2 ON u2.apMac = l2.apMac
GROUP BY u2.phoneMac) AS ijk
ON ijk.phoneMac = Users.phoneMac
AND Users.timestampfield = ij.latest;
(By the way, using the old join syntax with comma and the WHERE clause makes it harder to understand the logic, and occasionally makes the logic wrong. The new join syntax with ON is really a lot better.)