read_ef_file: Load a CSV file from the European Flux Database
In efreadr: Read European Eddy Fluxes CSV Files

Description Usage Arguments Details Value Note Examples

File name is parsed to extract year, site identification and aggregation type. The file name must point to a valid European Fluxes file, in CSV format and must resolve to a valid file format name.

1 2	read_ef_file(file_name, aggregation = NA_character_, year = NA_integer_, fill_value = -9999L)

`file_name`	Full path to 1 fluxes file
`aggregation`	character string, aggregation of the flux file to be imported (default `NA_character_`). This is important to properly define variable types for each aggregation type. When `NA_character_` it is assumed no aggregation (eg L2 files).
`year`	integer value (default `NA_integer`). Important to correctly add `efreadr_date` field
`fill_value`	a code for a not available (`NA`) observation in CSV file. All the observations with 'fill_value' values are converted to `NA`s during import. Default is -9999L.

File name is added as a field in the returned data frame as pathname. The pathname variable may be used to join a dataframe with file metadata such as year, site code, level, aggregation, as reaf_ef_files does.

The fluxes files code not available measures as -9999 (integer variables) or -9999.00 (floating point variables). Default behaviour of read_ef_file is to treat them as NAs. All -9999L, and -9999.0, -9999.00, -9999.000 values are converted to NA during import of files. Occasionally, -9999.00 (floting point representation of NA) appear in variables that are inherently integer (i.e. sqc variables in daily flux file). This forces a type conversion of the entire variable to double.

a data frame as loaded from the file, added with 'pathname' column, and 'efreadr_date' column for half-hourly fluxes files

For semi-hourly L4 aggregation (i.e. "h" aggregation in file name) the last row is reported as month 1, day 1, hour 00:00. A normal date conversion would convert this date to be the very first half-hour in January 1st of the current year whereas it should be the first half-hour of the January 1st of the following year. Therefore a class date field (efreader_date) is added to the returned data frame holding the correct date (ie: 1st January of the following year).