mirror of
https://github.com/asadbek064/hyparquet.git
synced 2025-12-28 07:56:38 +00:00
158 lines
4.2 KiB
JSON
158 lines
4.2 KiB
JSON
{
|
|
"version": 1,
|
|
"created_by": "parquet-mr version 1.8.1 (build 4aba4dae7bb0d4edbcf7923ae1339f28fd3f7fcf)",
|
|
"key_value_metadata": [
|
|
{
|
|
"key": "org.apache.spark.sql.parquet.row.metadata",
|
|
"value": "{\"type\":\"struct\",\"fields\":[{\"name\":\"a\",\"type\":\"string\",\"nullable\":true,\"metadata\":{}},{\"name\":\"b\",\"type\":\"integer\",\"nullable\":false,\"metadata\":{}},{\"name\":\"c\",\"type\":\"double\",\"nullable\":false,\"metadata\":{}},{\"name\":\"d\",\"type\":\"boolean\",\"nullable\":false,\"metadata\":{}},{\"name\":\"e\",\"type\":{\"type\":\"array\",\"elementType\":\"integer\",\"containsNull\":false},\"nullable\":true,\"metadata\":{}}]}"
|
|
}
|
|
],
|
|
"metadata_length": 836,
|
|
"num_rows": 5,
|
|
"row_groups": [
|
|
{
|
|
"columns": [
|
|
{
|
|
"file_offset": 4,
|
|
"meta_data": {
|
|
"codec": "SNAPPY",
|
|
"data_page_offset": 4,
|
|
"encodings": ["PLAIN", "RLE_DICTIONARY"],
|
|
"num_values": 5,
|
|
"path_in_schema": ["a"],
|
|
"statistics": {
|
|
"max": "abc",
|
|
"min": "abc",
|
|
"null_count": 1
|
|
},
|
|
"total_compressed_size": 63,
|
|
"total_uncompressed_size": 59,
|
|
"type": "BYTE_ARRAY"
|
|
}
|
|
},
|
|
{
|
|
"file_offset": 67,
|
|
"meta_data": {
|
|
"codec": "SNAPPY",
|
|
"data_page_offset": 67,
|
|
"encodings": ["DELTA_BINARY_PACKED"],
|
|
"num_values": 5,
|
|
"path_in_schema": ["b"],
|
|
"statistics": {
|
|
"max": 5,
|
|
"min": 1,
|
|
"null_count": 0
|
|
},
|
|
"total_compressed_size": 49,
|
|
"total_uncompressed_size": 47,
|
|
"type": "INT32"
|
|
}
|
|
},
|
|
{
|
|
"file_offset": 116,
|
|
"meta_data": {
|
|
"codec": "SNAPPY",
|
|
"data_page_offset": 116,
|
|
"encodings": ["PLAIN", "RLE_DICTIONARY"],
|
|
"num_values": 5,
|
|
"path_in_schema": ["c"],
|
|
"statistics": {
|
|
"max": 5,
|
|
"min": 2,
|
|
"null_count": 0
|
|
},
|
|
"total_compressed_size": 88,
|
|
"total_uncompressed_size": 94,
|
|
"type": "DOUBLE"
|
|
}
|
|
},
|
|
{
|
|
"file_offset": 204,
|
|
"meta_data": {
|
|
"codec": "SNAPPY",
|
|
"data_page_offset": 204,
|
|
"encodings": ["RLE"],
|
|
"num_values": 5,
|
|
"path_in_schema": ["d"],
|
|
"statistics": {
|
|
"max": true,
|
|
"min": false,
|
|
"null_count": 0
|
|
},
|
|
"total_compressed_size": 39,
|
|
"total_uncompressed_size": 37,
|
|
"type": "BOOLEAN"
|
|
}
|
|
},
|
|
{
|
|
"file_offset": 243,
|
|
"meta_data": {
|
|
"codec": "SNAPPY",
|
|
"data_page_offset": 243,
|
|
"encodings": ["PLAIN", "RLE_DICTIONARY"],
|
|
"num_values": 10,
|
|
"path_in_schema": [
|
|
"e",
|
|
"list",
|
|
"element"
|
|
],
|
|
"statistics": {
|
|
"max": 3,
|
|
"min": 1,
|
|
"null_count": 2
|
|
},
|
|
"total_compressed_size": 78,
|
|
"total_uncompressed_size": 74,
|
|
"type": "INT32"
|
|
}
|
|
}
|
|
],
|
|
"num_rows": 5,
|
|
"total_byte_size": 311
|
|
}
|
|
],
|
|
"schema": [
|
|
{
|
|
"name": "spark_schema",
|
|
"num_children": 5
|
|
},
|
|
{
|
|
"converted_type": "UTF8",
|
|
"name": "a",
|
|
"repetition_type": "OPTIONAL",
|
|
"type": "BYTE_ARRAY"
|
|
},
|
|
{
|
|
"name": "b",
|
|
"repetition_type": "REQUIRED",
|
|
"type": "INT32"
|
|
},
|
|
{
|
|
"name": "c",
|
|
"repetition_type": "REQUIRED",
|
|
"type": "DOUBLE"
|
|
},
|
|
{
|
|
"name": "d",
|
|
"repetition_type": "REQUIRED",
|
|
"type": "BOOLEAN"
|
|
},
|
|
{
|
|
"converted_type": "LIST",
|
|
"name": "e",
|
|
"num_children": 1,
|
|
"repetition_type": "OPTIONAL"
|
|
},
|
|
{
|
|
"name": "list",
|
|
"num_children": 1,
|
|
"repetition_type": "REPEATED"
|
|
},
|
|
{
|
|
"name": "element",
|
|
"repetition_type": "REQUIRED",
|
|
"type": "INT32"
|
|
}
|
|
]
|
|
}
|