Ask Search:
Evan PonterEvan Ponter 

Tips & Tricks: Count the Number of Date Groups in a Report

I've figured out a workaround for those interested in grouping report results by a date field and want to count the number of "groups" between the highest date value and the lowest date value, inclusive. This also includes any groupings in between those date values that aren't showing any records (desirable in some situations, not all. Just something to be aware of). While this isn't truly counting "the number of groups that you see in your report", it does work for a lot of situations where you have sequential groups of dates with data - e.g. records with January, February, and March dates on them and your report time frame is set to only show January, February, and March.

I will point out that Salesforce might be developing something to address this natively (https://www.salesforce.com/blog/2019/02/ideaexchange-formulas-dates.html). They mention that:
[The] team paired the two ideas together when scoping the feature, which helps report users organize and summarize data to answer questions such as:
  • When was the last time we contacted a customer?
  • How long, from the project start date, did it take to close an opportunity?
  • What’s the average number of records created within a given timespan?
The article goes on to say:
The initial beta functionality for the Summer ‘19 release will allow comparisons via date/time functions, and is currently in development for the Lightning Experience. Additional functions will be supported upon the general availability of the feature, planned for the Winter ‘20 release.

Which sounds like we should cross our fingers for a Winter '20 release. Until then, try this out:

You will need to create a formula field to correspond with the time frame for your group (see #1 below). This formula field assigns a number sequence to the record based on the date field. The starting number and actual values are arbitrary - this just needs to be a sequential series of numbers that are consistent for all records on the object so that you can calculate a difference between two records.

1. Create a formula field on the object in question.
Formula Field

Use the "number" data type with no decimal places.
Number Formula Options

Anything surrounded by brackets [ ] (including the brackets) needs to be replaced with your value.

- - - 1a. For fiscal year groupings, name the field "Fiscal Year of [date field name]" and use this formula. I'm using a fiscal-year-ending model here. You may want to use fiscal-year-beginning if your organization does so that there's no confusion (see below the code snippit), but it really doesn't matter since we are interested in the difference between two records. Also, I recommend storing your fiscal year start month and start day in custom settings so you don't hardcode those values here:
YEAR([date_field_API_name])
+
IF(
    MONTH([date_field_API_name]) > [Fiscal Year Start Month]
    ||
    (
        MONTH([date_field_API_name]) = [Fiscal Year Start Month]
        &&
        DAY([date_field_API_name]) >= [Fiscal Year Start Day]
    ),
    1,
/*ELSE*/
    0
)
For Fiscal Year Beginning, swap the plus sign for a minus sign on line 2, then swap the 1 and 0 on lines 10 and 12

- - - 1b. For calendar year groupings, name the field "Year of [date field name]" and use this formula:
YEAR([date_field_API_name])

- - - 1c. For calendar quarter groupings, name the field "Sequenced Quarter of [date field name]" and use this formula (works for fiscal quarters too as long as your fiscal months start on Jan 1, Apr 1, Jul 1, and Oct 1 - e.g. your fiscal year starts on Apr 1 and you have 3-month quarters resulting in fiscal months starting on Apr 1, Jul 1, Oct 1, and Jan 1):
IF(!ISBLANK([date_field_API_name]),
    CEILING(MONTH([date_field_API_name]) / 3)
    +
    4 * YEAR([date_field_API_name]),
/*ELSE*/
    NULL
)

- - - 1d. For calendar month groupings, name the field "Sequenced Month of [date field name]" and use this formula (works for fiscal months too as long as your fiscal months start on the 1st of each calendar month - e.g. Apr 1, May 1, Jun 1, etc):
IF(!ISBLANK([date_field_API_name]),
    MONTH([date_field_API_name])
    +
    12 * YEAR([date_field_API_name]),
/*ELSE*/
    NULL
)

- - - 1e. For calendar week groupings, name the field "Sequenced Week of [date field name]" and use this formula. This is for Sunday - Saturday weeks. To change which day of the week you want to start on, see below code snippit. I chose January 1, 1905 as my magic date because it's the beginning of a year pretty far in the past that falls on a Sunday. Even if your date value is before January 1, 1905, the calculations will still work since we are only concerned with the difference between two records:
IF(!ISBLANK([date_field_API_name]),
    FLOOR(([date_field_API_name] - DATE(1905,1,1)) / 7,
/*ELSE*/
    NULL
)
To start the week on Monday, replace line 2 with:
FLOOR(([date_field_API_name] - DATE(1905,1,2)) / 7,

To start the week on Tuesday, replace line 2 with:
FLOOR(([date_field_API_name] - DATE(1905,1,3)) / 7,

To start the week on Wednesday, replace line 2 with:
FLOOR(([date_field_API_name] - DATE(1905,1,4)) / 7,

To start the week on Thursday, replace line 2 with:
FLOOR(([date_field_API_name] - DATE(1905,1,5)) / 7,

To start the week on Friday, replace line 2 with:
FLOOR(([date_field_API_name] - DATE(1905,1,6)) / 7,

To start the week on Saturday, replace line 2 with:
FLOOR(([date_field_API_name] - DATE(1905,1,7)) / 7,

- - - 1f. For daily groupings, name the field "Sequenced Day of [date field name]" and use this formula:
[date_field_API_name] - DATE(1905,1,1)


2. In your report, create a custom summary formula.
To calculate the date groupings between the highest date and lowest date, specify the MAX of your formula field from #1 above, then subtract the MIN of your formula field from #1 above, then add 1 (since we want to count both the first and last group, not just what's in between).
[YourObjectHere].[YourNewSequenceFieldHere]:MAX
-
[YourObjectHere].[YourNewSequenceFieldHere]:MIN
+
1
Here's my example - I want to know how many "week" groups are in my report (Just for demo purposes to prove this method is calculating the right number):
Custom Summary Formula Counting Weeks
(this can be displayed at all levels or only at the grand total level)

Here's what it calculates when you run the report:
Report Showing Number of Weeks

Then here's a formula to calculate Average Number of Records per Week:
RowCount
/
(
    [YourObjectHere].[YourNewSequenceFieldHere]:MAX
    -
    [YourObjectHere].[YourNewSequenceFieldHere]:MIN
    +
    1
)
Display this at the grand total grouping only. You should see an accurate average calculation that takes into account how many date groupings you have in your report. Here's what mine looks like:
Average Records per Week

Hope this helps anyone out there trying to calculate something similar in their reports. Comment below with suggestions for improvement, or to celebrate when Salesforce builds this in natively.

Shout out to SteveMo for suggesting I post this. Also his tweet (https://twitter.com/SteveMoForce/status/1130904510833090560) was my motivation to buckling down on this endeavor
Best Answer chosen by Evan Ponter
Evan PonterEvan Ponter
Couple of improvements since I wrote this. Here are revised formulas for step 1. PLUS a few bonuses...


1a "Fiscal Year of [date field name]"
For consistency, this should only resolve to a value if there is a date in your date field.

I'm using a fiscal-year-ending model here. You may want to use fiscal-year-beginning if your organization does so that there's no confusion (see below the code snippit), but it really doesn't matter since we are interested in the difference between two records. Also, I recommend storing your fiscal year start month and start day in custom settings so you don't hardcode those values here:
IF(!ISBLANK([date_field_API_name]),
    YEAR([date_field_API_name])
    +
    IF(
        MONTH([date_field_API_name]) > [Fiscal Year Start Month]
        ||
        (
            MONTH([date_field_API_name]) = [Fiscal Year Start Month] 
            &&
            DAY([date_field_API_name]) >= [Fiscal Year Start Day]
        ),
        1,
    /*ELSE*/
        0
    ),
/*ELSE*/
    NULL
)
For Fiscal Year Beginning use:
IF(!ISBLANK([date_field_API_name]),
    YEAR([date_field_API_name])
    -
    IF(
        MONTH([date_field_API_name]) > [Fiscal Year Start Month]
        ||
        (
            MONTH([date_field_API_name]) = [Fiscal Year Start Month] 
            &&
            DAY([date_field_API_name]) >= [Fiscal Year Start Day]
        ),
        0,
    /*ELSE*/
        1
    ),
/*ELSE*/
    NULL
)


1b "Year of [date field name]"
For consistency, this should only resolve to a value if there is a date in your date field.
IF(!ISBLANK([date_field_API_name]),
    YEAR([date_field_API_name]),
/*ELSE*/
    NULL
)


1c. "Sequenced Quarter of [date field name]"
No changes needed. Same description from above:

For calendar quarter groupings, name the field "Sequenced Quarter of [date field name]" and use this formula (works for fiscal quarters too as long as your fiscal months start on Jan 1, Apr 1, Jul 1, and Oct 1 - e.g. your fiscal year starts on Apr 1 and you have 3-month quarters resulting in fiscal months starting on Apr 1, Jul 1, Oct 1, and Jan 1):
IF(!ISBLANK([date_field_API_name]),
    CEILING(MONTH([date_field_API_name]) / 3)
    +
    4 * YEAR([date_field_API_name]),
/*ELSE*/
    NULL
)


1d. "Sequenced Month of [date field name]"
No changes needed. Same description from above:

For calendar month groupings, name the field "Sequenced Month of [date field name]" and use this formula (works for fiscal months too as long as your fiscal months start on the 1st of each calendar month - e.g. Apr 1, May 1, Jun 1, etc):
IF(!ISBLANK([date_field_API_name]),
    MONTH([date_field_API_name])
    +
    12 * YEAR([date_field_API_name]),
/*ELSE*/
    NULL
)


1e "Sequenced Week of [date field name]"
This should really be using the MFLOOR function to handle dates before January 1, 1905. Plus my original post was missing a parenthesis on line 2 (oops!!). Use this instead for a Sunday - Saturday week:
IF(!ISBLANK([date_field_API_name]),
    MFLOOR(([date_field_API_name] - DATE(1905,1,1)) / 7),
/*ELSE*/
    NULL
)
Here's line 2 for a Monday - Sunday week:
MFLOOR(([date_field_API_name] - DATE(1905,1,2)) / 7),
Here's line 2 for a Tuesday - Monday week:
MFLOOR(([date_field_API_name] - DATE(1905,1,3)) / 7),
Here's line 2 for a Wednesday - Tuesday week:
MFLOOR(([date_field_API_name] - DATE(1905,1,4)) / 7),
Here's line 2 for a Thursday - Wednesday week:
MFLOOR(([date_field_API_name] - DATE(1905,1,5)) / 7),
Here's line 2 for a Friday - Thursday week:
MFLOOR(([date_field_API_name] - DATE(1905,1,6)) / 7),
Here's line 2 for a Saturday - Friday week:
MFLOOR(([date_field_API_name] - DATE(1905,1,7)) / 7),


1f "Sequenced Day of [date field name]"
For consistency, this should only resolve to a value if there is a date in your date field.
IF(!ISBLANK([date_field_API_name]),
    [date_field_API_name] - DATE(1905,1,1),
/*ELSE*/
    NULL
)


AND NOW INTRODUCING...
1g "Sequenced Business Day of [date field name]"

That's right, claiming the 1g spot is the "Sequenced Business Day of [date field name]" field. This will allow you to count the number of Business Days (Monday - Friday) in your report. You are still creating a number field with no decimal places, and you'd use this formula:
IF(!ISBLANK([date field name]) && WEEKDAY([date field name]) > 1 && WEEKDAY([date field name]) < 7,
    (MFLOOR(([date field name] - DATE(1905,1,1)) / 7) * 5)
    +
    WEEKDAY([date field name]),
/*ELSE*/
    NULL
)
Saturdays and Sundays will evaluate to a null value - which means they will be ignored completely even if they appear in your report. The weekdays will each be numbered in sequence. I checked this one for syntax, dates before 1/1/1905, etc so it should be an easy copy-and-paste for you. Then follow step 2 from the original post to create the custom summary formula in your report.


BONUS TIP 1:
You don't need to group your report results by the timeframe you are averaging. Confused? Here's my example. I am calculating the average number of records per week and I want to know that "weekly average" for a time frame spanning the whole fiscal year. I don't need to group my report results by week, only by fiscal year. The custom summary formula will take care of calculating the correct averge for each fiscal year, and the overall time frame in the total column
Average Per Week for Fiscal Years


BONUS TIP 2:
When using multiple groupings, you can use the custom summary formula as-is to find the difference between the maximum and minimum dates in each grouping, or use the PARENTGROUPVAL function to get the maximum and minimum dates from a higher grouping. This allows you to factor in the entire timeframe regardless of the data present in each grouping.

This example is for a report with row and column groupings (formerly known as a Matrix Report). This will need to be displayed at a grouping level (in my case Hostel/Community for the row and Start Date for the column).
RowCount
/
(
    PARENTGROUPVAL([YourObjectHere].[YourNewSequenceFieldHere]:MAX, ROW_GRAND_SUMMARY, [YourColumnGroupingFieldHere])
    -
    PARENTGROUPVAL([YourObjectHere].[YourNewSequenceFieldHere]:MIN, ROW_GRAND_SUMMARY, [YourColumnGroupingFieldHere])
    +
    1
)

There have been 9 weeks in FY 2020. Austin had a record each week, their average per week is 1. Nice.
Eastham had a record in each of the first 3 weeks, but none after that. The calculation for Eastham uses 3 as the numerator (the number of records for that group) and 9 as the demoninator (which is taken from the row grand summary for that column) and results in 0.33. For the completed years, each hostel is getting an average calculated based on a 52-week year regardless of having records in the first and last (or any) week of that year.
Average per Week for Entire Time Frame

All Answers

BhavinBhavin (salesforce.com) 
This is AWESOME, Evan! Thank you for sharing! :)

Best,
Bhavin
Evan PonterEvan Ponter
Couple of improvements since I wrote this. Here are revised formulas for step 1. PLUS a few bonuses...


1a "Fiscal Year of [date field name]"
For consistency, this should only resolve to a value if there is a date in your date field.

I'm using a fiscal-year-ending model here. You may want to use fiscal-year-beginning if your organization does so that there's no confusion (see below the code snippit), but it really doesn't matter since we are interested in the difference between two records. Also, I recommend storing your fiscal year start month and start day in custom settings so you don't hardcode those values here:
IF(!ISBLANK([date_field_API_name]),
    YEAR([date_field_API_name])
    +
    IF(
        MONTH([date_field_API_name]) > [Fiscal Year Start Month]
        ||
        (
            MONTH([date_field_API_name]) = [Fiscal Year Start Month] 
            &&
            DAY([date_field_API_name]) >= [Fiscal Year Start Day]
        ),
        1,
    /*ELSE*/
        0
    ),
/*ELSE*/
    NULL
)
For Fiscal Year Beginning use:
IF(!ISBLANK([date_field_API_name]),
    YEAR([date_field_API_name])
    -
    IF(
        MONTH([date_field_API_name]) > [Fiscal Year Start Month]
        ||
        (
            MONTH([date_field_API_name]) = [Fiscal Year Start Month] 
            &&
            DAY([date_field_API_name]) >= [Fiscal Year Start Day]
        ),
        0,
    /*ELSE*/
        1
    ),
/*ELSE*/
    NULL
)


1b "Year of [date field name]"
For consistency, this should only resolve to a value if there is a date in your date field.
IF(!ISBLANK([date_field_API_name]),
    YEAR([date_field_API_name]),
/*ELSE*/
    NULL
)


1c. "Sequenced Quarter of [date field name]"
No changes needed. Same description from above:

For calendar quarter groupings, name the field "Sequenced Quarter of [date field name]" and use this formula (works for fiscal quarters too as long as your fiscal months start on Jan 1, Apr 1, Jul 1, and Oct 1 - e.g. your fiscal year starts on Apr 1 and you have 3-month quarters resulting in fiscal months starting on Apr 1, Jul 1, Oct 1, and Jan 1):
IF(!ISBLANK([date_field_API_name]),
    CEILING(MONTH([date_field_API_name]) / 3)
    +
    4 * YEAR([date_field_API_name]),
/*ELSE*/
    NULL
)


1d. "Sequenced Month of [date field name]"
No changes needed. Same description from above:

For calendar month groupings, name the field "Sequenced Month of [date field name]" and use this formula (works for fiscal months too as long as your fiscal months start on the 1st of each calendar month - e.g. Apr 1, May 1, Jun 1, etc):
IF(!ISBLANK([date_field_API_name]),
    MONTH([date_field_API_name])
    +
    12 * YEAR([date_field_API_name]),
/*ELSE*/
    NULL
)


1e "Sequenced Week of [date field name]"
This should really be using the MFLOOR function to handle dates before January 1, 1905. Plus my original post was missing a parenthesis on line 2 (oops!!). Use this instead for a Sunday - Saturday week:
IF(!ISBLANK([date_field_API_name]),
    MFLOOR(([date_field_API_name] - DATE(1905,1,1)) / 7),
/*ELSE*/
    NULL
)
Here's line 2 for a Monday - Sunday week:
MFLOOR(([date_field_API_name] - DATE(1905,1,2)) / 7),
Here's line 2 for a Tuesday - Monday week:
MFLOOR(([date_field_API_name] - DATE(1905,1,3)) / 7),
Here's line 2 for a Wednesday - Tuesday week:
MFLOOR(([date_field_API_name] - DATE(1905,1,4)) / 7),
Here's line 2 for a Thursday - Wednesday week:
MFLOOR(([date_field_API_name] - DATE(1905,1,5)) / 7),
Here's line 2 for a Friday - Thursday week:
MFLOOR(([date_field_API_name] - DATE(1905,1,6)) / 7),
Here's line 2 for a Saturday - Friday week:
MFLOOR(([date_field_API_name] - DATE(1905,1,7)) / 7),


1f "Sequenced Day of [date field name]"
For consistency, this should only resolve to a value if there is a date in your date field.
IF(!ISBLANK([date_field_API_name]),
    [date_field_API_name] - DATE(1905,1,1),
/*ELSE*/
    NULL
)


AND NOW INTRODUCING...
1g "Sequenced Business Day of [date field name]"

That's right, claiming the 1g spot is the "Sequenced Business Day of [date field name]" field. This will allow you to count the number of Business Days (Monday - Friday) in your report. You are still creating a number field with no decimal places, and you'd use this formula:
IF(!ISBLANK([date field name]) && WEEKDAY([date field name]) > 1 && WEEKDAY([date field name]) < 7,
    (MFLOOR(([date field name] - DATE(1905,1,1)) / 7) * 5)
    +
    WEEKDAY([date field name]),
/*ELSE*/
    NULL
)
Saturdays and Sundays will evaluate to a null value - which means they will be ignored completely even if they appear in your report. The weekdays will each be numbered in sequence. I checked this one for syntax, dates before 1/1/1905, etc so it should be an easy copy-and-paste for you. Then follow step 2 from the original post to create the custom summary formula in your report.


BONUS TIP 1:
You don't need to group your report results by the timeframe you are averaging. Confused? Here's my example. I am calculating the average number of records per week and I want to know that "weekly average" for a time frame spanning the whole fiscal year. I don't need to group my report results by week, only by fiscal year. The custom summary formula will take care of calculating the correct averge for each fiscal year, and the overall time frame in the total column
Average Per Week for Fiscal Years


BONUS TIP 2:
When using multiple groupings, you can use the custom summary formula as-is to find the difference between the maximum and minimum dates in each grouping, or use the PARENTGROUPVAL function to get the maximum and minimum dates from a higher grouping. This allows you to factor in the entire timeframe regardless of the data present in each grouping.

This example is for a report with row and column groupings (formerly known as a Matrix Report). This will need to be displayed at a grouping level (in my case Hostel/Community for the row and Start Date for the column).
RowCount
/
(
    PARENTGROUPVAL([YourObjectHere].[YourNewSequenceFieldHere]:MAX, ROW_GRAND_SUMMARY, [YourColumnGroupingFieldHere])
    -
    PARENTGROUPVAL([YourObjectHere].[YourNewSequenceFieldHere]:MIN, ROW_GRAND_SUMMARY, [YourColumnGroupingFieldHere])
    +
    1
)

There have been 9 weeks in FY 2020. Austin had a record each week, their average per week is 1. Nice.
Eastham had a record in each of the first 3 weeks, but none after that. The calculation for Eastham uses 3 as the numerator (the number of records for that group) and 9 as the demoninator (which is taken from the row grand summary for that column) and results in 0.33. For the completed years, each hostel is getting an average calculated based on a 52-week year regardless of having records in the first and last (or any) week of that year.
Average per Week for Entire Time Frame
This was selected as the best answer
Ana FAna F
Hi Evan, 
This post is gold! I find it extremely insightful. 

There is a project that I am trying to work on that might benefit of this, but I am hoping to get some help from you on it. 
I am trying to calculate the average of calls (Activities) a sales rep had by week, but excluding the days in which he/she was out of office. 
I considered creating a custom Object to save the Absences, but how will I relate it to Activities? On the User level I cannot have a rollup summary, as no related lists are permitted.. 
Do you happen to have a recommendation for this please? 
Thank you!
Evan PonterEvan Ponter
Maybe we can get you close, here's what I'm thinking:

To clarify, you want to calculate the average number of calls per day and then have the report show you that calculation on a weekly basis for each sales rep? For example, if a rep has the following number of activities:
Monday: 1
Tuesday: 2
Wednesday: 3
Thursday: 4
Friday: 5
Then you would see an average of 3 calls per day for this rep for that week.

And then in this situation, you only want to count the number of days the rep actually worked:
Monday: 1
Tuesday: 2
Wednesday: absent
Thursday: 4
Friday: 5
So the average number of calls per day for this week would be (5+4+2+1)/4 = 3

You could try storing the absences in a custom object and be sure to relate them to a user (you could make use of the Owner field if this "Absences" object is not the detail object in a master-detail relationship). Then create the "Sequenced Day of Activity Date" on the Activity object as specified in 1f above, or if you just want Monday - Friday then set up a "Sequenced Business Day of Activity Date" field as-in 1g. You will also need to set up the same logic in a formula field on the Absences object.

Then you would have to use a joined report with one block set up for Activities and the other block set up for Absences. Group by user. You won't be able to do a sub-level grouping for the weeks across blocks in a joined report, so this report is going to have to be filtered down to a single week in each block. Hope that's not a show stopper?

In a cross-block custom summary formula, you can use a formula like this (the API names will likely be different in your org)
Activity_Block:RowCount
/
(
    Activity_Block.[YourNewSequenceFieldHere]:MAX
    -
    Activity_Block.[YourNewSequenceFieldHere]:MIN
    +
    1
    -
    (
        Absence_Block:RowCount
        -
        IF(Absence_Block:[YourNewSequenceFieldHere]:MAX = Activity_Block.[YourNewSequenceFieldHere]:MAX,
            1,
        /*ELSE*/
            0
        )
        -
        IF(Absence_Block:[YourNewSequenceFieldHere]:MIN = Activity_Block.[YourNewSequenceFieldHere]:MIN,
            1,
        /*ELSE*/
            0
        )
    )
)

The end result should assume each rep works the full week, but if they have any absences then that day will be subtracted from the calculation. If I have this worked out right, an absence at the beginning or ending of the time period you are filtering down to shouldn't throw off the calculation. I'm assuming there won't be any activities for a rep on the days they are absent, so a Friday absence would already calculate Monday as the min and Thursday as the max days with activites and we won't need to subtract that Friday absence record. A Wednesday and Friday absence would still see Monday as the min and Thursday as the max days with activites, so we would just subtract 1 day to account for Wednesday.

Hope this makes sense and points you in the right direction! Sorry for the limitation with looking at a single week and using joined reports... I don't have any better ideas at the moment to allow for the flexibility with tracking absences.
Jared RileyJared Riley
I'm trying to use this and it seemed to be working for a bit, but now some of my SRFs for Number of Months are showing a value of 2, when I believve they should all be 1. Any idea what would be causing this?User-added image
Evan PonterEvan Ponter
What field are you grouping the columns by? Is it a date field that has the grouping set to show calendar months? Is it the same field that you are using in the summary formula calculation? Could you post a screenshot showing the report groupings and another screenshot showing the formula you used?
Jared RileyJared Riley
I am grouping by Account Owner and Account Name (I believe is necessary for what I am trying calculate, which is average # of records per account over this given time)

User-added image
User-added image
Evan PonterEvan Ponter
Your filters must be allowing for opportunity records that are created in multiple months. This might be desirable for the calculation you want. To find the average number of opportunities created each month for each account owned by each user, your summary formula should be updated to:
RowCount
/
(
    Opportunity.Sequenced_Month_of_Created_Date__c:MAX
    -
    Opportunity.Sequenced_Month_of_Created_Date__c:MIN
    +
    1
)

This should calculate the correct average for each account as you change the time frame filters in your report.
Evan PonterEvan Ponter
Yes, you can wrap any datetime field in a DATEVALUE() function in order to reference just the date portion. So you can use 
DATEVALUE(CreatedDate)
anywhere you need to reference a date field in the above formulas.