Upgrading pyarrow 0.3 to 0.4
Breaking0.3.0 → 0.4.1Breaking2New29Changed7Fixed45Other39
BreakingUpgrade verdict
2 breaking changes.
In plain terms
- GLib code that called the old open constructor must switch to new.
- Python code using the old class name must switch to the new one.
Must act 2 items need code or config changes
What you must change
Breaking2
Renames the GLib constructor from open to new.
Renames InMemoryOutputStream to BufferOutputStream.
InMemoryOutputStreamBufferOutputStreamAPI changes
Generated by DepDelta from published Python source. Not mixed with prose notes.
- removed:
export FileReader - removed:
export FileWriter - removed:
export InMemoryOutputStream - removed:
export StreamReader - removed:
export StreamWriter - removed:
export pyarrow - added:
export BufferOutputStream - added:
export RecordBatchFileReader - added:
export RecordBatchFileWriter - added:
export RecordBatchStreamReader - added:
export RecordBatchStreamWriter - added:
export deserialize_pandas - added:
export open_file - added:
export open_stream - added:
export serialize_pandas - added:
variable FileReader - added:
variable FileWriter - added:
variable InMemoryOutputStream - added:
variable StreamReader - added:
variable StreamWriter
- changed:
function read_feather(source, columns=…) → function read_feather(source, columns=…, nthreads=…) - changed:
method FeatherReader.read(columns=…) → method FeatherReader.read(columns=…, nthreads=…) - added:
variable infer_dtype
- changed:
class HdfsClient(io._HdfsClient, Filesystem) → class HdfsClient(lib._HdfsClient, Filesystem) - added:
property Filesystem.pathsep - added:
property LocalFilesystem.pathsep
- removed:
class FileReader(_io._FileReader) - removed:
class FileWriter(_io._FileWriter) - removed:
class StreamReader(_io._StreamReader) - removed:
class StreamWriter(_io._StreamWriter) - removed:
constructor FileReader(source, footer_offset=…) - removed:
constructor FileWriter(sink, schema) - removed:
constructor StreamReader(source) - removed:
constructor StreamWriter(sink, schema) - added:
class RecordBatchFileReader(lib._RecordBatchFileReader) - added:
class RecordBatchFileWriter(lib._RecordBatchFileWriter) - added:
class RecordBatchStreamReader(lib._RecordBatchReader) - added:
class RecordBatchStreamWriter(lib._RecordBatchWriter) - added:
constructor RecordBatchFileReader(source, footer_offset=…) - added:
constructor RecordBatchFileWriter(sink, schema) - added:
constructor RecordBatchStreamReader(source) - added:
constructor RecordBatchStreamWriter(sink, schema) - added:
function deserialize_pandas(buf, nthreads=…) - added:
function open_file(source, footer_offset=…) - added:
function open_stream(source) - added:
function serialize_pandas(df)
- added:
function construct_metadata(df, index_levels, preserve_index) - added:
function get_column_metadata(column, name) - added:
function index_level_name(index, i) - added:
function infer_dtype(column) - added:
function is_unnamed_index_level(name) - added:
function table_to_blockmanager(table, nthreads=…) - added:
variable INDEX_LEVEL_NAME_REGEX
- added:
function read_pandas(source, columns=…, nthreads=…, metadata=…) - added:
method ParquetFile.read_pandas(columns=…, nthreads=…)