Skip to content

Optimize to_pandas() internally to improve performance #390

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
sethmlarson merged 5 commits into from
Oct 13, 2021
Merged

Optimize to_pandas() internally to improve performance #390

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
bfbe2ac
Optimize to_pandas() internally
V1NAY8 Sep 16, 2021
098f65f
Update operations.py
sethmlarson Oct 13, 2021
bb8672c
Handle the n=0 case for pandas.concat()
sethmlarson Oct 13, 2021
d7ce4a7
Update operations.py
sethmlarson Oct 13, 2021
596462d
Update query_compiler.py
sethmlarson Oct 13, 2021
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
62 changes: 22 additions & 40 deletions eland/operations.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -18,6 +18,7 @@
import copy
import warnings
from collections import defaultdict
from datetime import datetime
from typing import (
TYPE_CHECKING,
Any,
Expand All @@ -39,6 +40,7 @@
from eland.common import (
DEFAULT_PAGINATION_SIZE,
DEFAULT_PIT_KEEP_ALIVE,
DEFAULT_PROGRESS_REPORTING_NUM_ROWS,
DEFAULT_SEARCH_SIZE,
SortOrder,
build_pd_series,
Expand Down Expand Up @@ -1198,18 +1200,34 @@ def describe(self, query_compiler: "QueryCompiler") -> pd.DataFrame:
def to_pandas(
self, query_compiler: "QueryCompiler", show_progress: bool = False
) -> pd.DataFrame:
df = self._es_results(query_compiler, show_progress)

return df
df_list: List[pd.DataFrame] = []
i = 0
for df in self.search_yield_pandas_dataframes(query_compiler=query_compiler):
if show_progress:
i = i + df.shape[0]
if i % DEFAULT_PROGRESS_REPORTING_NUM_ROWS == 0:
print(f"{datetime.now()}: read {i} rows")
df_list.append(df)

if show_progress:
print(f"{datetime.now()}: read {i} rows")

# pd.concat() can't handle an empty list
# because there aren't defined columns.
if not df_list:
return query_compiler._empty_pd_ef()
return pd.concat(df_list)

def to_csv(
self,
query_compiler: "QueryCompiler",
show_progress: bool = False,
**kwargs: Union[bool, str],
) -> Optional[str]:
df = self._es_results(query_compiler, show_progress)
return df.to_csv(**kwargs) # type: ignore[no-any-return]
return self.to_pandas( # type: ignore[no-any-return]
query_compiler=query_compiler, show_progress=show_progress
).to_csv(**kwargs)

def search_yield_pandas_dataframes(
self, query_compiler: "QueryCompiler"
Expand Down Expand Up @@ -1241,42 +1259,6 @@ def search_yield_pandas_dataframes(
df = self._apply_df_post_processing(df, post_processing)
yield df

def _es_results(
self, query_compiler: "QueryCompiler", show_progress: bool = False
) -> pd.DataFrame:
query_params, post_processing = self._resolve_tasks(query_compiler)

result_size, sort_params = Operations._query_params_to_size_and_sort(
query_params
)

script_fields = query_params.script_fields
query = Query(query_params.query)

body = query.to_search_body()
if script_fields is not None:
body["script_fields"] = script_fields

# Only return requested field_names and add them to body
_source = query_compiler.get_field_names(include_scripted_fields=False)
body["_source"] = _source if _source else False

if sort_params:
body["sort"] = [sort_params]

es_results: List[Dict[str, Any]] = sum(
_search_yield_hits(
query_compiler=query_compiler, body=body, max_number_of_hits=result_size
),
[],
)

df = query_compiler._es_results_to_pandas(
results=es_results, show_progress=show_progress
)
df = self._apply_df_post_processing(df, post_processing)
return df

def index_count(self, query_compiler: "QueryCompiler", field: str) -> int:
# field is the index field so count values
query_params, post_processing = self._resolve_tasks(query_compiler)
Expand Down
17 changes: 2 additions & 15 deletions eland/query_compiler.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -16,7 +16,6 @@
# under the License.

import copy
from datetime import datetime
from typing import (
TYPE_CHECKING,
Any,
Expand All @@ -33,11 +32,7 @@
import numpy as np
import pandas as pd # type: ignore

from eland.common import (
DEFAULT_PROGRESS_REPORTING_NUM_ROWS,
elasticsearch_date_to_pandas_date,
ensure_es_client,
)
from eland.common import elasticsearch_date_to_pandas_date, ensure_es_client
from eland.field_mappings import FieldMappings
from eland.filter import BooleanFilter, QueryFilter
from eland.index import Index
Expand Down Expand Up @@ -149,7 +144,6 @@ def es_dtypes(self) -> pd.Series:
def _es_results_to_pandas(
self,
results: List[Dict[str, Any]],
show_progress: bool = False,
) -> "pd.Dataframe":
"""
Parameters
Expand Down Expand Up @@ -274,10 +268,6 @@ def _es_results_to_pandas(
# flatten row to map correctly to 2D DataFrame
rows.append(self._flatten_dict(row, field_mapping_cache))

if show_progress:
if i % DEFAULT_PROGRESS_REPORTING_NUM_ROWS == 0:
print(f"{datetime.now()}: read {i} rows")

# Create pandas DataFrame
df = pd.DataFrame(data=rows, index=index)

Expand All @@ -299,9 +289,6 @@ def _es_results_to_pandas(
if len(self.columns) > 1:
df = df[self.columns]

if show_progress:
print(f"{datetime.now()}: read {i} rows")

return df

def _flatten_dict(self, y, field_mapping_cache: "FieldMappingCache"):
Expand Down Expand Up @@ -383,7 +370,7 @@ def _index_matches_count(self, items: List[Any]) -> int:
self, self.index.es_index_field, items
)

def _empty_pd_ef(self):
def _empty_pd_ef(self) -> "pd.DataFrame":
# Return an empty dataframe with correct columns and dtypes
df = pd.DataFrame()
for c, d in zip(self.dtypes.index, self.dtypes.values):
Expand Down