Partition by and count

  • I'm trying to write a query that will count (N) all members within a specific city but without repeating the name of the city (N) times... any ideas to what is semi-working now?:

    SELECT city, 
    COUNT(PeopleID) OVER(PARTITION BY city) AS CountOfCities
    FROM gcdf_DB..PeopleContactInfo
  • SELECT city, 
    COUNT(*)
    FROM gcdf_DB..PeopleContactInfo
    GROUP BY City

    SELECT city,
    COUNT(DISTINCT PeopleID)
    FROM gcdf_DB..PeopleContactInfo
    GROUP BY City

    One of these two, depending on whether or not a PeopleID can appear multiple times against one city.

    “Write the query the simplest way. If through testing it becomes clear that the performance is inadequate, consider alternative query forms.” - Gail Shaw

    For fast, accurate and documented assistance in answering your questions, please read this article.
    Understanding and using APPLY, (I) and (II) Paul White
    Hidden RBAR: Triangular Joins / The "Numbers" or "Tally" Table: What it is and how it replaces a loop Jeff Moden

  • Is there a way to order by COUNT like?:

    New York 73,152

    Chicago 34,152

    Los Angeles 15,252

    I'm using this as a learning exercise too, so if anyone has any PARTITION BY or WITH CUBE blocks of code that would teach me - greatly appreciated

  • Sure

    GROUP BY City ORDER BY COUNT(*)

    “Write the query the simplest way. If through testing it becomes clear that the performance is inadequate, consider alternative query forms.” - Gail Shaw

    For fast, accurate and documented assistance in answering your questions, please read this article.
    Understanding and using APPLY, (I) and (II) Paul White
    Hidden RBAR: Triangular Joins / The "Numbers" or "Tally" Table: What it is and how it replaces a loop Jeff Moden

  • I'm trying to sum the payments for each person and I also wonder if I can take it further joining other tables like location, etc. and using SUM, ROW NUMBER and PARTITION BY to extract data like payments by state:

    My raw data looks like this:

    1 120.00 CO

    2 570.00 BC

    3 215.00 FL

    4 280.00 NC

    5 1190.00 IA

    SELECT distinct p.peopleId, SUM(p.Amount), s.stateAbbrveation
    FROM
    BCC_DB..payments p WITH(NOLOCK)
    INNER JOIN
    BCC_DB..peopleContactInfo i WITH(NOLOCK)
    ON p.peopleid = i.peopleid
    INNER JOIN BCC_DB..states s WITH(NOLOCK)
    ON i.stateid = s.stateid
    group by p.peopleId, s.stateAbbrveation
    ORDER BY p.peopleID

    • This reply was modified 4 years, 10 months ago by  DaveBriCam.
    • This reply was modified 4 years, 10 months ago by  DaveBriCam.
    • This reply was modified 4 years, 10 months ago by  DaveBriCam.
    • This reply was modified 4 years, 10 months ago by  DaveBriCam.
  • Please provide sample data in the form of a script to either create a temporary table or declare a table variable and an INSERT statement to insert values into said table.  Your sample should also contain at least on person who has more than one payment and at least one state that has more than one person.  We can't tell if the SUMs are working correctly if you only have one possible record for each group.

    You also need to provide expected results by, again, providing a temp table/table variable and inserted values.

    Drew

    J. Drew Allen
    Business Intelligence Analyst
    Philadelphia, PA

  • To add on to drew's comments you probably want to be clear about which state you want.  it looks like for now you're lumping it in by state of residence, not necessarily by state of point of sale.

    As to your query - if you're leveraging the GROUP BY syntax, you typically do not need or want a DISTINCT on the column list as well, since the group by will provide at most one entry per unique combination of the GROUP BY columns.  (it already handles the distinct for you).  it can also lead to some monstrously confusing results if the columns of the group by aren't included in the SELECT statement.

    ----------------------------------------------------------------------------------
    Your lack of planning does not constitute an emergency on my part...unless you're my manager...or a director and above...or a really loud-spoken end-user..All right - what was my emergency again?

Viewing 7 posts - 1 through 6 (of 6 total)

You must be logged in to reply to this topic. Login to reply