Character string truncated when passed to function - r

Question

I have a long list of Campaign names that I need to collapse to a character vector of length 1 and then pass as the "where" clause in a call to the Google AdWords API through the "RAdwords" package.
Creating this character string is not a problem until its length gets to a certain point that the values are truncated, which causes an error in AdWords API call.
Here is a sample of the setup that will not cause an error:
campaigns <- paste0("Campaign ", seq(1,5))
collapsed_campaigns <- paste0(campaigns, collapse = "','")
campaign_filter1 <- paste("CampaignName IN ['", collapsed_campaigns, "']")
And here is a setup that will cause an error:
campaigns <- paste0("Campaign ", seq(1,50))
collapsed_campaigns <- paste0(campaigns, collapse = "','")
campaign_filter2 <- paste("CampaignName IN ['", collapsed_campaigns, "']")
Inspecting the structure of each variable shows:
> str(campaign_filter1)
chr "CampaignName IN [' Campaign 1','Campaign 2','Campaign 3',
'Campaign 4','Campaign 5 ']"
> str(campaign_filter2)
chr "CampaignName IN [' Campaign 1','Campaign 2','Campaign 3',
'Campaign 4','Campaign 5','Campaign 6','Campaign 7','Campaign 8','Camp"| __truncated__
If I pass 'campaign_filter1' as my where clause in RAdwords, things run as expected.
If I pass 'campaign_filter2' as the where clause, I get this error:
<?xml version="1.0" encoding="UTF-8" standalone="yes"?><reportDownloadError>
<ApiError><type>QueryError.INVALID_WHERE_CLAUSE</type><trigger></trigger>
<fieldPath></fieldPath></ApiError></reportDownloadError>
It seems the "| truncated" is getting passed literally to the RAdwords function.
Here is the result of inspecting the structure of "traffic_data" in a failed call to RAdwords:
> str(traffic_data)
Classes ‘data.table’ and 'data.frame': 1 obs. of 1 variable:
$ ads: chr "<?xml version=\"1.0\" encoding=\"UTF-8\" standalone=\"yes\"?>
<reportDownloadError><ApiError><type>QueryError.INVALID_WHERE_CLAU"| __truncated__
- attr(*, ".internal.selfref")=<externalptr>
Obviously, I could get around this some sort of looping function and call the data from the API one campaign at a time, but that would be horribly inefficient. How can I get the entirety of the character string to be passed to RAdwords?

Answer 1

One question upfront: Why don´t you donwload all campaign data and filter the result dataframe in R? With this strategy the whole string campaign name pasting process would become superfluous. You could filter the dataframe based on vector operations in R. This approach probably is more rubust and less vulnerable.
However, if you want to filter campaigns explicitly in your API call you can do it with this code:
# 1. Download all campaigns
# query all campaign names
body1 <- statement(select=c('CampaignName'),
report="CAMPAIGN_PERFORMANCE_REPORT",
start="2017-11-01",
end="2017-11-02")
# download all campaign names
campaigns <- getData(clientCustomerId = "***-***-****",
google_auth = google_auth,
statement = body,
apiVersion = "201710",
transformation = T,
changeNames = T)
# 2. Build query with all campaigns in where clause
# build string for where clause
cmp_string <- paste0(campaigns$Campaign, collapse = "','")
cmp_string <- paste("CampaignName IN ['", cmp_string, "']", sep = "")
# query all campaigns with where condition
body2 <- statement(select = c('CampaignName'),
where = cmp_string,
report = "CAMPAIGN_PERFORMANCE_REPORT",
start = "2017-11-01",
end = "2017-11-02")
# download all campaigns using the where clause
campaigns2 <- getData(clientCustomerId = "***-***-****",
google_auth = google_auth,
statement = body,
apiVersion = "201710",
transformation = T,
changeNames = T)
In the first part I download all campaign names to have data for the where clause. In the second part I demonstrate how to download all campaigns again utilizing the where clause with all campaigns as filter.
I tested the code above with over 200 campaigns. Neither there were any issues with the RAdwords package nor with the Adwords API.
I suspect there are issues with the string you pass into campaign_filter2. Within paste() you miss to set sep = "". Otherwise you end up with a space in the beginning of the first campaign name.