Currently it's possible to run:

app_earnings amazon EarningsReport-Mar-1-2014-Mar-31-2014.csv PaymentReport-Mar-1-2014-April-4-2014.csv

And get ouput that seems plausible. However, there are no payments in the PaymentReport that match dates in the EarningsReport. Please fail with a message saying that the dates don't match (and what dates they are) as this is a point of frequent confusion. E.g. an error like:

Incompatible data ranges. The file EarningsReport-Mar-1-2014-Mar-31-2014.csv summary has dates from 3/1/2014,3/31/2014. The PaymentReport-Mar-1-2014-April-4-2014.csv has dates from 2/1/2014,2/28/2014.

This should be easy to detect using the Summary section of the Payment report.

• Earn money online app

In my latest EarningsReport, I have a refund amount in GBP, like this:

Amazon.co.uk,3/1/2014,3/31/2014,"1234",-2,99.99,GBP,(1.67),GBP,0.00,GBP,98.32,GBP

The final amount in my report is off by the exactly the amount of 1.67 GBP when converted at the exchange rate given in the corresponding PaymentReport. It appears that the refunds column is being ignored completely in this scenario and hence the total doesn't match at the end. (Interestingly, perhaps, there are 2 refunds mentioned in the report details, but again, perhaps they aren't being calculated? Not sure ...)

As of Feb/March 2017, Google Play Payment reports are missing the Product id and Sku id fields. This is only for lines where the sales report code starts with "GPA".

The lack of these fields makes generating Google Play reports either fail outright, or not attribute all sales correctly.

I've worked around this as follows:

1. Create a list of product ids & sku ids from a previous play report as follows:
   a. install csvkit - https://csvkit.readthedocs.io/en/749/
   b. run the following command on a previous play report file that has all possible products and skus:

grep -v GPA PlayApps_xxxxxx.csv | 
csvcut -c 7,8,10 | 
sort -u | 
grep -v "Product Title" | 
sed -e '1s/^/Product Title,Product id,Sku id\'$'\n/' >map-names.csv

2. Hack the source code of app_earnings/google_play/parse.rb so it reads as follows:

require 'csv'

module AppEarnings::GooglePlay
  # Converts a csv file to a hash.
  class Parser
    attr_accessor :file_name

    def initialize(file_name)
      load_fixup
      @file_name = file_name
    end

    def extract
      @extracted_data = []
      options = { headers: true, header_converters: :symbol }

      CSV.foreach(@file_name, options) do |row|
        hash_row = row.to_hash
        hash_row = do_fixup(hash_row) if hash_row[:product_id].nil?
        @extracted_data << hash_row
      end
      @extracted_data
    end

    def load_fixup
      @fixup_table = []
      options = { headers: true, header_converters: :symbol }

      CSV.foreach("map-names.csv", options) do |row|
        @fixup_table << row.to_hash
      end
    end

    def do_fixup(row)
      fix = @fixup_table.select {|f| f[:product_title] == row[:product_title]}[0]
      row[:product_id] = fix[:product_id]
      row[:sku_id] = fix[:sku_id]
      row
    end
  end
end

That's it.

Hopefully Google stop producing files like this. Otherwise, I'll make an effort to clean this up and merge it into the code in a more elegant fashion.