NEWS

datapackage 0.2.3 (2025-05-28)

Under windows incorrect line endings were written: '\r\r\n' instead of '\r\n' (issue #25). Fixed.
The 'use_fwrite' argument of 'csv_writer' was not passed on to lower functions. Therefore, 'fwrite' was not used when this argument was set to 'TRUE' (issue #26). Fixed.
‘print.fielddescriptor' could give an error for field with categories defined, when those categories do not use the default names ’value' and 'label'. This has been fixed.
When saving a ‘datapackage.json' file all length 1 vectors were written as scalar values. However, some properties should be written as length 1 vectors according to the data package specification. For these fields this is now the case. For custom fields vectors of length 1 are still ’unboxed'.

‘dp_apply_schema()' now handles the 'fieldsMatch' property of the schema correctly. Previously only '’exact'' was handled: the columns in the data set should match the fields in the schema exactly.
'dp_check_dataresource()' was looking for the the 'fieldsMatch' property in the wrong location in the schema (it expected this property to be directly in the Data Resource). This property should be in the 'schema' property of the Data Resource.
'csv_reader()' now handles the 'fieldsMatch' property of the schema correctly. Previously the the code assumed 'exact' for the 'fieldsMatch' property. In general this would result in an error when reading a file.
'dp_apply_schema()' gives clearer error messages: reports the column in which the error occured.
Small changes to print functions.
Fix issue with large integer when using a yaml file. These are now converted to numeric.
'dp_write_data()' gave error when data is stored in a subdirectory of the data package and this directory does not yet exist. These directories are now automatically created.

'dp_to_code' will remove white space at the beginning and end of character values from category values and the field and multiple hypens at the beginning of values when the option 'DP_TRIM_CODES' is set to 'TRUE' (default value is 'FALSE'). Ignoring multiple hyphens can be disabled by setting 'DP_TRIM_HYPHEN' to 'FALSE'.
'dp_to_factor' will ignore white space at the beginning and end of character values and multiple hypens at the beginning of values when the option 'DP_TRIM_CODES' is set to 'TRUE' (default value is 'FALSE'). Ignoring multiple hyphens can be disabled by setting 'DP_TRIM_HYPHEN' to 'FALSE'.
'dp_generate_fielddescriptor' now generates correct meta data for objects of type 'code' from the 'codelist' package.
'dp_categorieslist' returns the code list from a 'code' object from the 'codelist' package.
'dp_type()' and 'dp_type<-()' added for Field Descriptor to get and set the field type.
Added checking of constraints for fields of type 'datetime', 'year', 'yearmonth'.

Checking 'minimum', 'maximum', 'exclusiveMinimum', 'exclusiveMaximum' constraints on non numeric fields failed.