There are incorrect results in the response, in particular for TX DC120149D1C33A4AADA230698DB06355F7890E02785420AAD2E210BEA0F37914 first item should be in USD (currency code and amount), yet somehow it became XRP (both currency code and amount)
Also there is a problem with TX 9C5A288843EE896A882F12D006D0EEA4605785E40755B160759CF074B516B3BB. There is an additional "exchange" item created for XRP exchange side with node_index=null (the correct one seems to have node_index correctly specified)
DataAPI request "exchanges" also seems to have duplicate items for that time interval.
Please fix the issues, accurate data is very important for accounting.
Check out a DataAPI balance_changes request like this: https://data.ripple.com/v2/accounts/rEp4SYVfwQgGbWCJV4tBAgGNpG2KeiaY1W/balance_changes?descending=false&limit=1000&start=2017-09-12T10:58:42Z&end=2017-09-12T11:00:10Z&change_type=exchange
There are incorrect results in the response, in particular for TX DC120149D1C33A4AADA230698DB06355F7890E02785420AAD2E210BEA0F37914 first item should be in USD (currency code and amount), yet somehow it became XRP (both currency code and amount)
Also there is a problem with TX 9C5A288843EE896A882F12D006D0EEA4605785E40755B160759CF074B516B3BB. There is an additional "exchange" item created for XRP exchange side with node_index=null (the correct one seems to have node_index correctly specified)
Those incorrect results seem to be concentrated on days 2017-09-11 and 2017-09-12 (UTC time). There are many of them in that time interval. There are even exchange items created for XRP when no actual exchange happened (seems to be created for every offer, even though it did not result in any exchanges ) with node_index=null as well. Check request like this for that: https://data.ripple.com/v2/accounts/rEp4SYVfwQgGbWCJV4tBAgGNpG2KeiaY1W/balance_changes?descending=false&limit=1000&start=2017-09-11T00:00:00Z&end=2017-09-13T00:00:00Z&change_type=exchange
DataAPI request "exchanges" also seems to have duplicate items for that time interval. Please fix the issues, accurate data is very important for accounting.