Note that the four (4) .csv files listed below, and that appears in each "date" folder contain errors.  It is preferable to use the NETCDF format for a complete dataset. 
Noter que les quatre (4) fichiers .csv listés ci-dessous et qui apparaissent dans chacun des dossiers "date" contiennent des erreurs.  Il est préférable d'utiliser le format NETCDF pour un ensemble de données complet.  


1. All of the time fields need to be converted to 64bit float, not 32bit. At 32bit, the mjd times are truncated to the nearest couple of minutes which means certain adjacent scans will report they are at the same time. This is a problem as many times analysis is done using time as an index and the indices must be unique. This applies to the following files in all of the CSV data product folders:
* Geolocation Fields/Time.csv
* Geolocation Fields/ScanStartTime.csv
* Geolocation Fields/ScanEndTime.csv

2. The Scan numbers (which are the unique id for the OSIRIS scans recorded as orbit*1000 + scan, e.g. 6432012 for scan 12 of orbit 6432) are also recorded as 32bit float in the csv files and should be recorded as an integer (a 32 bit integer is sufficient). As it is now, when converted back to integer, all scans with an orbit of 10000 or less come back missing the last digit in the scan number and those in orbits after 10000 come back missing the last two digits of the scan number. This applies to the following file type in all of the CSV data product folders:
* Geolocation Fields/ScanNo.csv