i am using unionall function to get the combined results of multiple queries but i get duplicate data ,
sample example :
select expand($res) let $a = (select from #3:1) , $b = (select from #3:1) , $res = unionall($a,$b)
i get duplicate data ,
even tried by querying rid , but still same result
select $res let $a = (select rid from #3:1) , $b = (select rid from #3:1) ,$res = unionall($a,$b)
how do i get unique values ?
select distinct(#this)
from (select expand($u) as res
from (select $u
let $a = (select from #3:1),
$b = (select from #3:1),
$u = unionall($a,$b)) )
Note that "select rid from #3:1" is quite different from "select #rid from #3:1", which in turn is different from "select from #3:1".
Try this query
select expand($res) let $a = (select from #3:1) , $b = (select from #3:1) , $res = intersect($a,$b)
I replaced unionall with intersect.
Related
$query = "select (select count(a.serial_no) from tra2 a where a.model_no = ".$id." and a.flag = cast(3 as character varying) + (select count(a.serial_no) from stk a where a.model_no = ".$id." and a.trans_id is NULL)as qty)";
$result = $this->db->query($query);
return $result->result();
When i run the query
ERROR: syntax error at or near "as"
LINE 1: ...ck a where a.model_no = K258 and a.trans_id is NULL)as qty)
Add '' (quotes) on your WHERE clauses. I guess the data type is VARCHAR so you have to use quotes in your query
Like this
.... WHERE a.model_no = '" . $id . "' AND ...
The end of the quest must be:
...a.trans_id is NULL)) as qty"
Currently You have some like this
select (select 1 + 2 as qty)
but You want
select (select 1 + 2) as qty
I want the result to return the max Rank when partitioned using the rank function.
I am using the following query.
SELECT DISTINCT dbo.pomst.co_num
,dbo.pomst.wh_num
,dbo.pomst.po_number
,dbo.pomst.po_suffix
,dbo.pomst.vendor_id
,dbo.item.uom
,dbo.item.upc_num
,dbo.item.item_desc
,RIGHT(dbo.auditlog.pallet_id, 8) AS pallet_id
,dbo.auditlog.abs_num
,dbo.auditlog.item_qty
,dbo.auditlog.lot
,dbo.auditlog.packer
,auditlog.comments
,auditlog.date_time
,rank() OVER (
PARTITION BY auditlog.comments ORDER BY auditlog.date_time ASC
) AS CorrectTrans
FROM dbo.auditlog
INNER JOIN dbo.pomst ON dbo.auditlog.co_num = dbo.pomst.co_num
AND dbo.auditlog.wh_num = dbo.pomst.wh_num
AND dbo.auditlog.po_number = dbo.pomst.po_number
AND dbo.auditlog.po_suffix = dbo.pomst.po_suffix
INNER JOIN dbo.item ON dbo.auditlog.co_num = dbo.item.co_num
AND dbo.auditlog.wh_num = dbo.item.wh_num
AND dbo.auditlog.abs_num = dbo.item.abs_num
WHERE (dbo.pomst.co_num = 'AC01')
AND (dbo.pomst.wh_num = 'KU22')
AND (dbo.pomst.row_status = 'C')
AND (dbo.auditlog.trans_type = 're')
AND item_qty NOT LIKE '-%'
I figured it out! I was trying to get the max result of a rank, but if I flip order of rank from asc to desc and use a CTE I can select the the results that always have 1 as the rank as opposed to trying to get the Max. I would still like to know how to get the Max rank but this solution suits my needs.
;with cte as
(SELECT DISTINCT dbo.pomst.co_num
,dbo.pomst.wh_num
,dbo.pomst.po_number
,dbo.pomst.po_suffix
,dbo.pomst.vendor_id
,dbo.item.uom
,dbo.item.upc_num
,dbo.item.item_desc
,RIGHT(dbo.auditlog.pallet_id, 8) AS pallet_id
,dbo.auditlog.abs_num
,dbo.auditlog.item_qty
,dbo.auditlog.lot
,dbo.auditlog.packer
,auditlog.comments
,auditlog.date_time
,rank() OVER (
PARTITION BY auditlog.comments ORDER BY auditlog.date_time desc
) AS CorrectTrans
FROM dbo.auditlog
INNER JOIN dbo.pomst ON dbo.auditlog.co_num = dbo.pomst.co_num
AND dbo.auditlog.wh_num = dbo.pomst.wh_num
AND dbo.auditlog.po_number = dbo.pomst.po_number
AND dbo.auditlog.po_suffix = dbo.pomst.po_suffix
INNER JOIN dbo.item ON dbo.auditlog.co_num = dbo.item.co_num
AND dbo.auditlog.wh_num = dbo.item.wh_num
AND dbo.auditlog.abs_num = dbo.item.abs_num
WHERE (dbo.pomst.co_num = 'AC01')
AND (dbo.pomst.wh_num = 'KU22')
AND (dbo.pomst.row_status = 'C')
AND (dbo.auditlog.trans_type = 're')
AND item_qty NOT LIKE '-%'
)
Select * from cte
where CorrectTrans = 1
Add select and group by and use your existing query as a sub query.
Try ..
select max([CorrectTrans]), Vendor_Id, Item_qty, Lot, Pallet_id
from (
-- Your existing query --
SELECT DISTINCT dbo.pomst.co_num
,dbo.pomst.wh_num
,dbo.pomst.po_number
,dbo.pomst.po_suffix
,dbo.pomst.vendor_id
,dbo.item.uom
,dbo.item.upc_num
,dbo.item.item_desc
,RIGHT(dbo.auditlog.pallet_id, 8) AS pallet_id
,dbo.auditlog.abs_num
,dbo.auditlog.item_qty
,dbo.auditlog.lot
,dbo.auditlog.packer
,auditlog.comments
,auditlog.date_time
,rank() OVER (
PARTITION BY auditlog.comments ORDER BY auditlog.date_time ASC
) AS CorrectTrans
FROM dbo.auditlog
INNER JOIN dbo.pomst ON dbo.auditlog.co_num = dbo.pomst.co_num
AND dbo.auditlog.wh_num = dbo.pomst.wh_num
AND dbo.auditlog.po_number = dbo.pomst.po_number
AND dbo.auditlog.po_suffix = dbo.pomst.po_suffix
INNER JOIN dbo.item ON dbo.auditlog.co_num = dbo.item.co_num
AND dbo.auditlog.wh_num = dbo.item.wh_num
AND dbo.auditlog.abs_num = dbo.item.abs_num
WHERE (dbo.pomst.co_num = 'AC01')
AND (dbo.pomst.wh_num = 'KU22')
AND (dbo.pomst.row_status = 'C')
AND (dbo.auditlog.trans_type = 're')
AND item_qty NOT LIKE '-%'
-- =======================================
) x
group by Vendor_id, Item_qty, Lot, Pallet_id
I get error (The multi-part identifier "f.FormID" could not be bound.) running this query:
select f.FormID, f.Title, fv.UserName
from Forms f join (
SELECT FormID
FROM Reports
WHERE (ReportID = #ReportID)
UNION
SELECT FormRelations.ForigenFormID
FROM FormRelations INNER JOIN
Forms ON FormRelations.ForigenFormID = Forms.FormID
WHERE (FormRelations.PrimaryFormID =
(SELECT FormID
FROM Reports
WHERE (ReportID = #ReportID)))
) ids
on f.FormID = ids.FormID
LEFT OUTER JOIN (select top 1 UserName, FormID from FormValues where FormID = f.FormID and UserName = #UserName) fv
ON f.FormID = fv.FormID
Please someone help me :(
#bluefeet:
I want such a result:
01304636-FABE-4A3E-9487-A14B012F9A61 item_1 1234567890
C0455E97-788A-4305-876A-A15000CFE928 item_2 1234567890
7719F37E-7021-4ABD-91ED-A15301830324 item_3 1234567890
If you need to use your alias inside of your subquery like that, you might want to look at using the APPLY operator:
select f.FormID, f.Title, fv.UserName
from Forms f
join
(
SELECT FormID
FROM Reports
WHERE (ReportID = #ReportID)
UNION
SELECT FormRelations.ForigenFormID
FROM FormRelations
INNER JOIN Forms
ON FormRelations.ForigenFormID = Forms.FormID
WHERE (FormRelations.PrimaryFormID = (SELECT FormID
FROM Reports
WHERE (ReportID = #ReportID)))
) ids
on f.FormID = ids.FormID
CROSS APPLY
(
select top 1 UserName, FormID
from FormValues
where FormID = f.FormID
and UserName = #UserName
) fv
Or you can use row_number():
select f.FormID, f.Title, fv.UserName
from Forms f
join
(
SELECT FormID
FROM Reports
WHERE (ReportID = #ReportID)
UNION
SELECT FormRelations.ForigenFormID
FROM FormRelations
INNER JOIN Forms
ON FormRelations.ForigenFormID = Forms.FormID
WHERE (FormRelations.PrimaryFormID = (SELECT FormID
FROM Reports
WHERE (ReportID = #ReportID)))
) ids
on f.FormID = ids.FormID
LEFT JOIN
(
select UserName, FormID,
ROW_NUMBER() over(PARTITION by FormID, UserName order by FormID) rn
from FormValues
where UserName = #UserName
) fv
on f.FormID = fv.FormID
and fv.rn = 1
trying to add a conditional AND.
I don't want to check for the Data field match if the #Data has nothing
select distinct mo.ModelId,
mo.ModelName
from Model mo
join MaModelar mmy on mmy.ModelID = mo. ModelID
join [Application] a on a.MaModelar ID = mmy.MaModelar ID
join Appltitle t on t.ApplTID = a.ApplicationID
where mmy.MakeId = 331
CASE
WHEN LEN(t.Name) > 0
THEN
and (t.Name = 'A.T.V.')
END
order by mo.ModelName asc
Let me summarize in different words:
I don't want to query on (Modelid AND #Note) if #Note is invalid. Invalid means #Note is empty or null. So I am trying to add some kind of conditional statement that first checks #Note and if it's valid include it in the where clause as an AND. So if #Note is valid (len > 1) then this query should include mmy.MakeId = 331 AND t.Data = 'A.T.V.'.
If #Note is not valid, then the query would just have WHERE mmy.MakeId = 331 and leave out the AND and it would not try to check for #Note along with mmy.MakeId.
I tried this also:
select distinct mo.ModelId,
mo.ModelName
from Model mo
join MaModelar mmy on mmy.ModelID = mo. ModelID
join [Application] a on a.MaModelar ID = mmy.MaModelar ID
join Appltitle t on t.ApplTID = a.ApplicationID
WHEN LEN(LTRIM(RTRIM(#Name))) < 1
THEN where mmy.ModelId = #ModelId
ELSE
where mmy.ModelId = #ModelId
AND substring(n.Name, CHARINDEX(']', n.Name) + 2, LEN(n.Name)) = #Name
use
select distinct mo.ModelId,
mo.ModelName
from Model mo
join MaModelar mmy on mmy.ModelID = mo. ModelID
join [Application] a on a.MaModelar ID = mmy.MaModelar ID
join Appltitle t on t.ApplTID = a.ApplicationID
where mmy.MakeId = 331 AND ( LEN(t.Data) < 1 OR t.Data = 'A.T.V.')
order by mo.ModelName asc
EDIT - as per comments:
try
select distinct mo.ModelId,
mo.ModelName
from Model mo
join MaModelar mmy on mmy.ModelID = mo. ModelID
join [Application] a on a.MaModelar ID = mmy.MaModelar ID
join Appltitle t on t.ApplTID = a.ApplicationID
where mmy.MakeId = 331 AND t.Data = 'A.T.V.'
order by mo.ModelName asc
Just have a nested filter:
WHERE mmy.MakeID = 331
AND (T.Data = '' OR t.Data = 'A.T.V.')
Using the LEN function will make the query non-SARGable so I would avoid that if possible.
Just change your and clause to
AND
(substring(n.Name, CHARINDEX(']', n.Name) + 2, LEN(n.Name)) = #Name
or #Name is null)
Note: there's no way to force the DB Engine to short-circuit the OR clause.
I also do this in some cases
DECLARE #IsNameEmpty bit
IF #Name is null or #Name = '' or #Name = 'Some Special value'
#IsNameEmpty = 1
else
#IsNameEmpty = 0
and then I do this instead
AND
(substring(n.Name, CHARINDEX(']', n.Name) + 2, LEN(n.Name)) = #Name
or #IsNameEmpty = 1)
Which makes it easier if I have a query that uses #Name several times and I don't feel like writing each condition.
you don't need to worry about null fields. Your check t.Data = 'A.T.V.' will fail if the column is empty.
I want to place this working code within a SQL Statement, OR do I need to perform a UDF.
The result set is a one line concatenation, and I want it to be place in every one of the overall result set lines.
----
MAIN QUERY
SELECT
H.CONNECTION_ID,
H.SEQUENTIAL_NO,
H.INVOICE_NUMBER,
H.INVOICE_DATE,
H.LAST_INVOICE_NUMBER,
H.LAST_INVOICE_DATE,
CAST(CASE
WHEN H.COLLECT_DEPOSIT = 1 THEN '-'
ELSE CAST(H.PAYMENT_DUE_DATE AS NVARCHAR(20))
END AS SMALLDATETIME) AS PAYMENT_DUE,
H.JOB_NUMBER,
H.CUST_JOB_NUMBER,
HDR.SALES_PERSON,
H.INSIDE_SALES_PERSON,
H.IS_LAST_INVOICE,
CASE
WHEN H.COLLECT_DEPOSIT = 1 THEN 'CASH'
ELSE H.PAYMENT_TERMS_DESCRIPTION
END AS PAYMENT_TERMS,
H.PRINTED,
H.NOTES,
CUR.ID,
CUR.CODE,
CASE CUR.CODE
WHEN 'USD' THEN '001-106624-211'
WHEN 'EUR' THEN '001-106624-101'
WHEN 'GBP' THEN '001-106624-100'
ELSE '001-106624-001'
END AS BANK_ACCT,
CUR.EXCHANGE_RATE,
H.BILL_CONTACT,
H.CUST_ACCOUNT,
H.CUST_NAME,
H.CUST_ADDR1,
H.CUST_ADDR2,
H.CUST_CITY,
H.CUST_STATE,
H.CUST_ZIP,
H.CONTACT_PHONE_NUMBER,
H.CONTACT_PHONE_NUMBER2,
H.ORDERED_BY_CONTACT,
H.SHIP_TO_NAME,
H.SHIP_TO_ADDR1,
H.SHIP_TO_ADDR2,
H.SHIP_TO_CITY,
H.SHIP_TO_STATE,
H.SHIP_TO_ZIP,
H.SITE_PHONE_NUMBER,
H.SITE_PHONE_NUMBER2,
H.OFFICE_NAME,
H.OFFICE_ADDR1,
H.OFFICE_ADDR2,
H.OFFICE_CITY,
H.OFFICE_STATE,
H.OFFICE_ZIP,
H.OFFICE_PHONE_NUMBER,
H.OFFICE_FAX_NUMBER,
H.DELIVERY_TICKET_NUMBER,
H.PO_NUMBER,
H.DUMMY_INVOICE_TEXT,
(SELECT MESSAGE FROM REPORT_MESSAGES WHERE CODE = 'INVOICE') ADVERT_MESSAGE,
(SELECT MAX(DISCOUNT_PERCENTAGE) FROM PRTINVITEM I2 WHERE I2.CONNECTION_ID = H.CONNECTION_ID AND I2.INVOICE_NUMBER = H.INVOICE_NUMBER) AS MAX_DISCOUNT,
I.ITEM,
I.DESCRIPTION,
I.QUANTITY,
I.UNIT_OF_MEASURE,
I.MINIMUM_CHARGE,
I.WEEKLY_CHARGE,
I.MONTHLY_CHARGE,
I.START_OF_BILLING_PERIOD,
I.END_OF_BILLING_PERIOD,
I.DAYS_USED,
I.WEEKS_USED,
I.DISCOUNT_PERCENTAGE,
I.TAX_CODE_FOR_ITEM,
I.INVENTORY_TYPE,
I.BILLING_LOGIC_TYPE,
I.ACTUAL_WEEKLY_CHARGE_USED,
I.DAYS_IN_ACTUAL_WEEKLY_CHARGE,
II.CHARGEABLE_DAYS,
II.CHARGEABLE_WEEKS,
II.CHARGEABLE_MONTHS,
II.FREE_DAYS_THIS_INVOICE,
CNV.TOTAL_NET_VALUE,
CNV.TOTAL_TAX_VALUE,
CNV.TOTAL_GROSS_VALUE,
CNV.TOTAL_GROSS_VALUE_NS,
CNV.NET_LINE_VALUE,
CMP.EMAIL_ADDRESS
FROM (PRTINVHDR H INNER JOIN PRTINVITEM I ON H.CONNECTION_ID = I.CONNECTION_ID AND H.INVOICE_NUMBER = I.INVOICE_NUMBER)
INNER JOIN INVOICEHDR HDR ON I.INVOICE_NUMBER = HDR.INVNO
INNER JOIN CUSTOMERS CST ON H.CUST_ACCOUNT = CST.CUSTNUM
INNER JOIN JOB JOB ON H.JOB_NUMBER = JOB.JOBNUM
INNER JOIN CURRENCY CUR ON HDR.CURRENCY_ID = CUR.ID
INNER JOIN VWCURRENCYCONVERSION CNV ON I.CONNECTION_ID = CNV.CONNECTION_ID AND I.INVC_UCOUNTER = CNV.INVC_UCOUNTER
INNER JOIN COMPANY CMP ON H.OFFICE_CODE = CMP.OFFICE
INNER JOIN INVOICEITEM II ON I.INVOICE_NUMBER = II.INVNO AND I.INVC_UCOUNTER = II.INVC_UCOUNTER
ORDER BY
H.SEQUENTIAL_NO,
I.PRINT_SEQUENCE
ASC
----
COALESCE QUERY
DECLARE
#DTICKET NVARCHAR(20),
#PUMPCATEGORYNAME NVARCHAR(3999)
SET #DTICKET = ''
SET #PUMPCATEGORYNAME = NULL
(SELECT
#DTICKET = DTICKET,
#PUMPCATEGORYNAME = COALESCE(#PUMPCATEGORYNAME + ', ', '' ) + PUMPCATEGORYNAME
FROM (SELECT
BHDR.DTICKET,
SCD.PUMPCATEGORYNAME
FROM PRTTICKHDR PHDR
INNER JOIN BIDHDR BHDR ON PHDR.DELIV_TICKET_NUMBER = BHDR.DTICKET
INNER JOIN PRTTICKITEM PITM ON PHDR.CONNECTION_ID = PITM.CONNECTION_ID AND PHDR.DELIV_TICKET_NUMBER = PITM.DELIV_TICKET_NUMBER
LEFT JOIN SUBCATEGORYDESCRIPTION SCD ON PITM.ITEM = SCD.PUMPCATEGORY
WHERE SCD.PUMPCATEGORYNAME IS NOT NULL)
SUBCATEGORYDESCRIPTION)
SELECT #DTICKET, #PUMPCATEGORYNAME
Not really sure what you are asking for but you can doing something along the lines of
Select col1 + ', ' + col2 + ', ' + col3 etc....