forked from futurewei-cloud/caerus-dike
-
Notifications
You must be signed in to change notification settings - Fork 4
Commit
This commit does not belong to any branch on this repository, and may belong to a fork outside of the repository.
Merge pull request #6 from rf972/master
Parquet support
- Loading branch information
Showing
17 changed files
with
338 additions
and
53 deletions.
There are no files selected for viewing
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| Original file line number | Diff line number | Diff line change |
|---|---|---|
| @@ -0,0 +1,176 @@ | ||
| from __future__ import print_function | ||
|
|
||
| import argparse | ||
| import sys | ||
| import base64 | ||
|
|
||
| import pyarrow as pa | ||
| import pandas as pd | ||
| import pyarrow.parquet as pq | ||
|
|
||
|
|
||
| def get_metadata(parquet_file): | ||
| r = pq.ParquetFile(parquet_file) | ||
| return r.metadata | ||
|
|
||
| def get_row_group_metadata(parquet_file, group): | ||
| r = pq.ParquetFile(parquet_file) | ||
| return r.metadata.row_group(group) | ||
|
|
||
| def get_row_group_col_metadata(parquet_file, group, col): | ||
| r = pq.ParquetFile(parquet_file) | ||
| return r.metadata.row_group(group).column(col) | ||
|
|
||
| def print_key_value_metadata(parquet_file): | ||
| meta = pq.read_metadata(parquet_file) | ||
| #print("\n{}\n".format(meta.metadata.keys())) | ||
| for key in meta.metadata.keys(): | ||
| print("\n{}:\n {}".format(key, meta.metadata[key])) | ||
|
|
||
| #decoded_schema = base64.b64decode(meta.metadata[b"ARROW:schema"]) | ||
| #schema = pa.ipc.read_schema(pa.BufferReader(decoded_schema)) | ||
| #print(schema) | ||
|
|
||
| def get_schema(parquet_file): | ||
| r = pq.ParquetFile(parquet_file) | ||
| return r.schema | ||
|
|
||
|
|
||
| def get_data(pq_table, n, head=True): | ||
| data = pq_table.to_pandas() | ||
| if head: | ||
| rows = data.head(n) | ||
| else: | ||
| rows = data.tail(n) | ||
| return rows | ||
|
|
||
| def write_multiplied_table(pq_table, output_prefix): | ||
| df = pq_table.to_pandas() | ||
| add_multiplier = 1 | ||
| prev_df = df | ||
| for i in range(0, add_multiplier): | ||
| print("append {} ".format(i)) | ||
| new_df = prev_df.append(df) | ||
| prev_df = new_df | ||
| print("new_df rows {}".format(len(new_df.index))) | ||
| new_table = pa.Table.from_pandas(new_df) | ||
| pq.write_table(new_table, output_prefix + "ex_large.snappy.parquet", data_page_size=100*1024*1024*1024) | ||
|
|
||
| def write_table(pq_table, output_prefix): | ||
| df = pq_table.to_pandas() | ||
| rows = len(df.index) | ||
| print("rows {}".format(rows)) | ||
| print("output file " + output_prefix) | ||
| new_df = df[:int(7)] | ||
| new_table = pa.Table.from_pandas(new_df) | ||
| pq.write_table(new_table, output_prefix + ".parquet.snappy", data_page_size=1024*1024*1024) | ||
| #pq.write_table(pq_table, output_prefix + "full.snappy.parquet", data_page_size=1024*1024*1024) | ||
|
|
||
| def query_table(pq_table, query): | ||
| df = pq_table.to_pandas() | ||
| rows = len(df.index) | ||
| print("rows {}".format(rows)) | ||
| print("output: " + query) | ||
| new_df = df[['l_quantity','l_extendedprice','l_discount','l_shipdate']] | ||
| new_table = pa.Table.from_pandas(new_df) | ||
| pq.write_table(new_table, "query" + ".parquet.snappy", data_page_size=1024*1024*1024) | ||
|
|
||
| def main(cmd_args=sys.argv, skip=False): | ||
| """ | ||
| Main entry point with CLI arguments | ||
| :param cmd_args: args passed from CLI | ||
| :param skip: bool, whether to skip init_args call or not | ||
| :return: string stdout | ||
| """ | ||
| if not skip: | ||
| cmd_args = init_args() | ||
|
|
||
| pd.set_option('display.max_columns', None) | ||
| pd.set_option('display.max_rows', None) | ||
|
|
||
| pq_table = pq.read_table(cmd_args.file) | ||
| if cmd_args.write: | ||
| #write_multiplied_table(pq_table, cmd_args.write) | ||
| write_table(pq_table, cmd_args.write) | ||
| elif cmd_args.query: | ||
| query_table(pq_table, cmd_args.query) | ||
| elif cmd_args.head: | ||
| print(get_data(pq_table, cmd_args.head)) | ||
| elif cmd_args.tail: | ||
| print(get_data(pq_table, cmd_args.tail, head=False)) | ||
| elif cmd_args.count: | ||
| print(len(pq_table.to_pandas().index)) | ||
| elif cmd_args.schema: | ||
| print("\n # Schema \n", get_schema(cmd_args.file)) | ||
| else: | ||
| file_md = get_metadata(cmd_args.file) | ||
| print("\n # Metadata \n", file_md) | ||
| for i in range(0, file_md.num_row_groups): | ||
| rg_md = get_row_group_metadata(cmd_args.file, i) | ||
| print("\n # Row Group {}\n".format(i), rg_md) | ||
| if cmd_args.verbose: | ||
| for c in range(0, rg_md.num_columns): | ||
| col_md = get_row_group_col_metadata(cmd_args.file, i, c) | ||
| print("\n Column {}\n".format(c), col_md) | ||
|
|
||
| print_key_value_metadata(cmd_args.file) | ||
|
|
||
| def init_args(): | ||
| parser = argparse.ArgumentParser( | ||
| description="Command line (CLI) tool to inspect Apache Parquet files on the go", | ||
| usage="usage: parq file [-s [SCHEMA] | --head [HEAD] | --tail [TAIL] | -c [COUNT]]" | ||
| ) | ||
|
|
||
| group = parser.add_mutually_exclusive_group() | ||
|
|
||
| group.add_argument("-s", | ||
| "--schema", | ||
| nargs="?", | ||
| type=bool, | ||
| const=True, | ||
| help="get schema information", | ||
| ) | ||
|
|
||
| group.add_argument("--head", | ||
| nargs="?", | ||
| type=int, | ||
| const=10, | ||
| help="get first N rows from file" | ||
| ) | ||
|
|
||
| group.add_argument("--tail", | ||
| nargs="?", | ||
| type=int, | ||
| const=10, | ||
| help="get last N rows from file" | ||
| ) | ||
|
|
||
| group.add_argument("-c", | ||
| "--count", | ||
| nargs="?", | ||
| type=bool, | ||
| const=True, | ||
| help="get total rows count", | ||
| ) | ||
|
|
||
| group.add_argument("-w", "--write", default="", | ||
| help="output file") | ||
|
|
||
| group.add_argument("-q", "--query", default="", | ||
| help="run a query on a dataframe") | ||
|
|
||
| group.add_argument("-v", "--verbose", default=False, action="store_true", | ||
| help="more detailed output") | ||
|
|
||
| parser.add_argument("file", | ||
| type=argparse.FileType('rb'), | ||
| help='Parquet file') | ||
|
|
||
| cmd_args = parser.parse_args() | ||
|
|
||
| return cmd_args | ||
|
|
||
|
|
||
| if __name__ == '__main__': | ||
| args = init_args() | ||
| main(args, skip=True) |
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Oops, something went wrong.