Cohort example ignores first cohort

[borenstein]

Dear Mark,

First of all, thank you for writing this great R package! I am delighted to be able to use the API v4 directly from R.

I tried out the cohort example from the vignette, but it didn't behave the way I expected.

## first make a cohort group
cohort4 <- make_cohort_group(list("cohort 1" = c("2015-08-01", "2015-08-01"), 
                                "cohort 2" = c("2015-07-01","2015-07-01")))

## then call cohort report.  No date_range and must include metrics and dimensions
##   from the cohort list
cohort_example <- google_analytics_4(ga_id, 
                                     dimensions=c('cohort'), 
                                     cohort = cohort4, 
                                     metrics = c('cohortTotalUsers'))

This only returns a value for cohort 2. My first thought was that we didn't have data for cohort 1, but if I swapped the date ranges, I got only data for 2015-08-01. In fact, if I added a third row, I get the last two:

## first make a cohort group
cohort4 <- make_cohort_group(list("cohort 1" = c("2015-08-01", "2015-08-01"), 
                                "cohort 2" = c("2015-07-01","2015-07-01"),
                                "cohort 3" = c("2015-08-01", "2015-08-01")))

## then call cohort report.  No date_range and must include metrics and dimensions
##   from the cohort list
cohort_example <- google_analytics_4(ga_id, 
                                     dimensions=c('cohort'), 
                                     cohort = cohort4, 
                                     metrics = c('cohortTotalUsers'))

The above returns cohorts 2 and 3. Seems to be a simple off-by-one. I am using R 3.2.4 and I installed via install.packages today.

[borenstein]

It appears that the return object is leaving off the first row of the results, but it's running the whole query. So consider the following example:

cohort4 <- make_cohort_group(list("August 2015" = c("2015-08-01", "2015-08-31"), 
                                  "July 2015" = c("2015-07-01","2015-07-31")))

## then call cohort report.  No date_range and must include metrics and dimensions
##   from the cohort list
cohort_example <- google_analytics_4(ga_id, 
                                     dimensions=c('cohort', "ga:cohortNthMonth"), 
                                     cohort = cohort4, 
                                     metrics = c('cohortTotalUsers', "ga:cohortRetentionRate"))

This returns a 7x4 table. The August 2015 cohort is missing the 0-month row. So that's where the issue lies.

[MarkEdmondson1234]

Thanks very much for this and detective work, definitely a bug that I will look to fix ASAP.

[MarkEdmondson1234]

Well good news is that this is now fixed, bad news is that it was affecting all v4 calls, missing out the first row.

This was due to batching where I added the pageToken to the call, and in my R-brain I indexed it from 1, whereas it should have been from 0. I will submit this as a patch to the CRAN version ASAP.

Thanks very much for the report!

[borenstein]

Thank you for developing the package!
On May 24, 2016 6:07 PM, "Mark" notifications@github.com wrote:

Well good news is that this is now fixed, bad news is that it was
affecting all v4 calls, missing out the first row. This was due to batching
add the pageToken and in my R-brain I indexed it from 1, whereas it should
have been from 0. I will submit this as a patch to the CRAN version ASAP.

Thanks very much for the report!

—
You are receiving this because you authored the thread.
Reply to this email directly or view it on GitHub
https://github.com/MarkEdmondson1234/googleAnalyticsR/issues/16#issuecomment-221416071

[MarkEdmondson1234]

This is now on CRAN, version 0.1.1.