Closed jsta closed 5 years ago
Issue is that some fields in the exports are detected as integer even though they contain substantial non-integer values (with decimals). The end result is that about 20-30% of the srp and no2 data, which happens to be decimal, is dropped.
Are there special characters in _epinutr somewhere that is breaking read_csv
?
See this script for a method to manually give integers one decimal place so that readr
automatically detects as numeric (doubles): https://gist.github.com/jsta/070d9a70eb0850d552daf1856282db73
Confirmed fixed in LAGOSNE limno version 1.087.3 for all definitely numeric epinutr
data columns (i.e. not one of censorcode
, qual
, etc).
If the exports are downloaded and read into R (outside the LAGOSNE package) with the tidy underscore functions there are excessive NA cells possibly due to incorrect column type detection. Are there special characters responsible for the errors that could be fixed upstream?