DataFrameRField

Specific RField for Dataframe, these are loaded and dumped into a file.

DataFrames are dumped to Parquet (columnar, compressed, no code-execution risk on read). For backward compatibility, load_from_file still reads legacy pickle files: it sniffs the parquet magic bytes and falls back to read_pickle when they are absent, so DataFrames saved before this change keep loading.

A DataFrame that Parquet cannot represent (e.g. object columns holding arbitrary Python objects) falls back to pickle on dump, so a save never fails.

WARNING: pickle files (legacy files, and the rare fallback above) are read with pandas' pickle reader, which can execute arbitrary code. Only load files that you trust. DataFrames written as Parquet are not affected by this risk.

:param FileRField: [description] :type FileRField: [type]

Attributes

include_in_dict_view: boolstorage: RFieldStorage

Functions

__init__

Initialize a FileRField with automatic file-based persistence.

The field is configured to store file path references in KV_STORE and is excluded from dict views since the actual data is in files.

deserialize

Deserialize field value by loading data from the file path.

This method is marked as final and cannot be overridden. It delegates to the load_from_file method which must be implemented by subclasses.

r_field_value : str

File path where the data is stored

Return type : T

dump_to_file

Serialize and save the field value to a file.

This abstract method must be implemented by subclasses to define how data is serialized and written to the file. It is called automatically when the Resource is saved.

The data MUST be written to the exact file path provided. The directory containing the file is guaranteed to exist.

r_field_value : DataFrame

The field value to serialize and save

file_path : str

Absolute path where the data should be saved

get_default_value

Get the default value for this field.

If the default value is a Type or Callable, it will be called to generate a new default value. Otherwise, the default value is returned directly.

This ensures that mutable defaults (like lists or dicts) are not shared between Resource instances.

Return type : Any

load_from_file

Load and deserialize the field value from a file.

This abstract method must be implemented by subclasses to define how data is read and deserialized from the file. It is called automatically when the Resource is loaded.

The file at file_path is guaranteed to exist when this method is called.

file_path : str

Absolute path to the file containing the serialized data

Return type : DataFrame

serialize

Serialize method not used for FileRField.

The actual serialization is handled by dump_to_file which is called by the resource persistence layer. This method exists for interface compatibility but should be ignored.

r_field_value : Any

The field value (ignored)

Return type : str

Have you developed a brick?