I am new to MDX expressions and queries and I am currently looking for something like this -
I have two dates 'Date1' & 'Date2' in Fact Table as Foreign Keys to DATE dimension. I need to do a count of rows where Date1 < Date 2, keeping in account that I don't want to count NULLS. I wrote an expression something like this -
WITH MEMBER [Measures].[RecordCount] AS
COUNT(FILTER([Measures].[RecordCount], IIF([Date1].[Date] <= [Date2].[Date],0,1)=1))
SELECT [Measures].[RecordCount] ON 0
FROM [MYCUBE]
The above queries runs fine, but the count turns out to be incorrect. I created 7 rows in my fact table where Date1 is less than Date2, but still I receive the count as 0.
Any help is appreciated. (any reference sites would be good too for future)
Thanks,
Vineet
vineet9860#gmail.com
You can't really do this easily in MDX, the [RecordCount] measure will be aggregated up before you do the comparison of the dates, so you will never get a valid value.
You would be better to create a named calculation in your DSV that has something like
CASE WHEN Date1 < Date2 THEN 1 ELSE NULL END
and then create a measure in your cubes that sums up this new column.
Related
I have a table like this in SQLITE3:
I need to query this table by ID|DOC_ID|TRANS_DOC_ID and most importantly by DATE because I need to get the data day by day. ex: TODAY|YESTERDAY|ETC
So far the query is easy, as I can just do this to get the rows by day:
SELECT * FROM CLIENTRECORD WHERE DATE = '2020-12-01'
The problem is when I need to display specific records on other dates:
ex: I have a row with DATE 2020-12-01 but I also want it displayed on DATE 2020-01-01 or maybe 2020-01-02, etc. What do I do in this situation? and so I thought about adding another col as DATES which was supposed to be an array of comma-separated dates BUT I researched that this is a BAD solution, I also thought about adding a separate TABLE just for dates but since the dates aren't fixed (they might contain 1 date or maybe even 10 who knows), I am confused as to what I am supposed to do.
The end goal is that a row may or may not contain more than 1 date, would look something like this if I want to query for the row with or without multiple dates:
SELECT * FROM CLIENTRECORD WHERE DATE = '2020-12-01' OR DATES LIKE '2020-12-01'
something similar to it.
So I've been looking around Stack and couldn't find a solution for my particular need.
So currently I'm trying to count to see a Pat's check in to a Practice - here is my table and some notes on what I'm trying to achieve (FYI - trying to resolve the solution using T-SQL)
A check in is considered when a the same Pat ID and the same Practice are present and the dates are consecutive
Other wise its a new check in
So there is a grouping column that indicates that the PatID and the PracticeIDs are the same and is ordering by dates - but - I can't seem to wrap my head around trying to look at the previous EndDate and seeing that the next Row's start date is consecutive.
I'm thinking it'll need some sort of Recursion/Cursor stuff going on but I'm not entirely sure where to start.
All help appreciated!
The following query should only return rows from your table for PatID/PracticeID combinations for which there does NOT exist another row in your table for that same combination but with a DateEnd that is one date prior to the DateStart.
SELECT o.PatID, o.PracticeID
FROM <table> o
WHERE NOT EXISTS( SELECT 1
FROM <table> i
WHERE o.PatID = i.PatID
AND o.PracticeID = i.PracticeID
AND DATEDIFF(d, i.DateEnd, o.DateStart) = 1 )
GROUP BY o.PatID, o.PracticeID
Note: the order of i.DateEnd and o.DateStart may need to be reversed.
I have the following Access query (see above) to give me the date, X number of weeks out from the actual WorkDte that exists in the table. But, the date these expressions calculate out may not actually exists in the “CIB_Results” table due to a bank holiday etc. Is there a way to generate the same sort of data but say if it’s running the “Wk1” calculation and it calculates out 1/1/2016 (which does not exists due to the new year holiday), instead of doing [WorkDte]-7 it will move on to [WorkDte]-14 and so on, until it finds an actual date that exists in the “CIB_Results” table? I’m wanting to apply the same logic to all the fields in the query…that way they will all self adjust based off of the actual dates that exist in the “CIB_Results” table. Any help with this would be greatly appreciated!
Have a function IsHoliday(SomeDate) that looks up SomeDate in your holiday table and returns True if found.
Then create a loop:
Public Function PreviousWorkWeekday(ByVal SomeDate As Date) As Date
Do
SomeDate = DateAdd("ww", -1, SomeDate)
Loop Until Not IsHoliday(SomeDate)
PreviousWorkWeekday = SomeDate
End Function
Now use function PreviousWorkWeekday in your query.
I have a couple of tables in PowerPivot:
A Stock table - WKRelStrength whose fields are:
Ticker, Date, StockvsMarket% (values are percentages), RS+- (values can be 0 or 1)
A Calendar Table - Cal with a Date field.
There is a many to one relationship between the tables.
I am trying to aggregate RS+-against each row for dates between 3 months ago to the date for that row - i.e a 3 month to date sum. I have tried numerous calculations but the best I can return is an circular reference error. Here is my formula:
=calculate(sum([RS+-]),DATESINPERIOD(Cal[Date],LASTDATE(Cal[Date]),-3,Month))
Here is the xlsx file.
I couldn't download the file but what you are after is what Rob Collie calls the 'Greatest Formula in the World' (GFITW). This is untested but try:
= CALCULATE (
SUM ( WKRelStrength[RS+-] ),
FILTER (
ALL ( Cal ),
Cal[Date] <= MAX ( Cal[Date] )
&& Cal[Date]
>= MAX ( Cal[Date] ) - 90
) )
Note, this will give you the previous 90 days which is approx 3 months, getting exactly the prior 3 calendar months may be possible but arguably is less optimal as you are going to be comparing slightly different lengths of time (personal choice I guess).
Also, this will behave 'strangely' if you have a total in that it will use the last date in your selection.
First of all, the formula that you are using is designed to work as a Measure. This may not work well for a Calculated Column. Secondly, it is better to do such aggregations as a Measure level, than at individual records.
Then again, I do not fully understand your situation, but if it is absolutely important for you to do this at a Record level, you may want to use the "Earlier" Function.
If you want to filter a function, based on a value in the correspontinf row, you just have to wrap your Column name with the Earlier Function. Try changing the LastDate to Earlier in your formula.
I am new to SQL. I was given a coursework to report data of usage over the last 2 month. Can someone help me with the SQL statement?
SELECT COUNT(Member_ID,Non_Member_Name) AS Pool_usage_last_2_months
FROM Use_of_pool
WHERE DATEDIFF(‘2012-04-21’,’2012-02-21’)
What I meant to do is to count the total number of member usage(member_ID) and non member usage(no ID,name only) from the last two months and then output the name and date and time,etc. on the same report. Is there any SQL statement to output that kind of information? Correction/Suggestions are welcomed.
You need a different WHERE clause. Assuming your Use_of_pool table includes a Date/Time field, date_field:
WHERE date_field >= #2012-02-21# AND date_field <= #2012-04-21#
If date_field values can include a time component other than midnight, advance the end date range by one day to capture all the possible Date/Time values from Apr. 21:
WHERE date_field >= #2012-02-21# AND date_field <= #2012-04-22#
That should restrict the rows to match what I think you want. It should offer fast performance with an index on date_field.
I'm unclear about the count(s) you want ... whether it is to be one count for all visits (both member and non-member), or separate counts for members and non-members.
Edit: If each row of the table represents a visit by one person, you can simply count the rows to determine the number of visits during your selected time frame.
SELECT Count(*) AS CountOfVisits
FROM Use_of_pool
WHERE date_field >= #2012-02-21# AND date_field <= #2012-04-21#
Notice each visit by the same person will contribute to CountOfVisits, which is what I think you want. If you wanted to know how many different people visited, we will need a different approach.
Edit2:
It sounds like you can use Member_ID and Non_Member_Name to distinguish between member and nonmember visits. Member_ID is Null for nonmembers and non-Null for members. And Non_Member_Name is Null for members and non-Null for nonmembers.
If that is true, try this query to count member and nonmember visits separately.
SELECT
Sum(IIf(Member_ID Is Not Null, 1, 0)) AS member_visits,
Sum(IIf(Non_Member_Name Is Not Null, 1, 0)) AS non_member_visits
FROM Use_of_pool
WHERE date_field >= #2012-02-21# AND date_field <= #2012-04-21#
Aggregate functions of SQL use all the data in a column (more precisely, all the data your WHERE clause selects) to produce a single datum. COUNT gives you the number of data rows that matched your WHERE clause. So for example:
SELECT COUNT(*) AS Non_members FROM Use_of_pool WHERE Member_ID IS NULL
will give you the number of times the pool was used by a non-member, and
SELECT COUNT(DISTINCT Member_ID) AS Members FROM Use_of_pool
will give you the number of members who have used the pool at least once (the DISTINCT tells the database engine to ignore duplicates when counting).
You can expand the WHERE clause to further specify what you want to count. If "last two months" means the current and previous calendar month, you'll need:
... WHERE DateDiff("m",Date_field,Date())<=1
If it means a rolling 2-month period, I'd approximate that with 60 days and say
... WHERE DateDiff("d",Date_field,Date())<60
(Replace Date_field with the name of the field containing the date.)
If you want to count rows according to multiple different criteria, or output both aggregate data and individual data, you'll be best off using separate SELECT statements.