github.com/fraugster/parquet-go@v0.12.0/TODO.md (about)

     1  # Open TODOs
     2  
     3  * add functionality to help with managing schema evolution (forward- and backwards-compatibility).
     4  * add test for type store implementations to check whether the min and max values are correctly tracked
     5  * verify whether blockSize: 128 and miniBlockCount in (\*byteArrayDeltaLengthEncoder).Close() is correct.
     6  * in (\*byteArrayStore).setMinMax() whether the bytes.Compare calls are correct.
     7  * rewrite booleanPlainEncoder implementation using packed array.
     8  * readPageData: having a dictEncoder/decoder is wrong. they should be a plain decoder for header and a int32 hybrid for values. the mix should happen here not in the dict itself
     9  * writeChunk: check whether parquet.Encoding\_RLE is actually required.
    10  * improve (\*ColumnStore).reset() so that it works without losing schema information in the typed column store.
    11  * check whether (\*FileWriter).FlushRowGroup() should still return an error if the number of records in the row group is 0.
    12  * in (\*FileWriter).FlushRowGroup() add support for sorting columns.
    13  * in (\*FileWriter).Close() add support for column orders.
    14  * check whether it is feasible to implement a block cache in the packed array implementation
    15  * dictPageWriter: add support for sorted dictionary.
    16  * schema.go: the current design suggest every reader is only on one chunk and its not concurrent support. we can use multiple reader but its better to add concurrency support to the file reader itself
    17  * schema.go: add validation so every parent at least have one child.
    18  * (\*schema).ensureRoot(): a hacky way to make sure the root is not nil (because of my wrong assumption of the root element) at the last minute. fix it
    19  * (\*schema).ensureRoot(): provide a way to override the root column name
    20  * parquet-tool cat: add support for detailed schema (-d)
    21  * parquet-tool head: add support for detailed schema (-d)
    22  * parquet-tool schema: add support for detailed schema (-d)