yujunjun/datahub
2
0
Fork 0
mirror of https://github.com/datahub-project/datahub.git synced 2025-11-04 04:39:10 +00:00
Code Issues Packages Projects Releases Wiki Activity

fix(profiling): don't stop if some steps fail (#5095)

Browse Source
This commit is contained in:
Aseem Bansal 2022-06-07 03:06:32 +05:30 committed by GitHub
parent 64c4f518a4
commit c43ad54a64
No known key found for this signature in database
GPG Key ID: 4AEE18F83AFDEB23
1 changed files with 55 additions and 6 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

61
metadata-ingestion/src/datahub/ingestion/source/ge_data_profiler.py
View File

@ -305,8 +305,17 @@ class _SingleDatasetProfiler(BasicDatasetProfilerBase):
def _get_column_cardinality( def _get_column_cardinality(
self, column_spec: _SingleColumnSpec, column: str self, column_spec: _SingleColumnSpec, column: str
) -> None: ) -> None:
nonnull_count = self.dataset.get_column_nonnull_count(column) try:
column_spec.nonnull_count = nonnull_count nonnull_count = self.dataset.get_column_nonnull_count(column)
column_spec.nonnull_count = nonnull_count
except Exception as e:
logger.debug(
f"Caught exception while attempting to get column cardinality for column {column}. {e}"
)
self.report.report_warning(
"Profiling - Unable to get column cardinality",
f"{self.dataset_name}.{column}",
)
unique_count = None unique_count = None
pct_unique = None pct_unique = None
@ -352,21 +361,43 @@ class _SingleDatasetProfiler(BasicDatasetProfilerBase):
def _get_dataset_column_median( def _get_dataset_column_median(
self, column_profile: DatasetFieldProfileClass, column: str self, column_profile: DatasetFieldProfileClass, column: str
) -> None: ) -> None:
if self.config.include_field_median_value: if not self.config.include_field_median_value:
return
try:
column_profile.median = str(self.dataset.get_column_median(column)) column_profile.median = str(self.dataset.get_column_median(column))
except Exception as e:
logger.debug(
f"Caught exception while attempting to get column median for column {column}. {e}"
)
self.report.report_warning(
"Profiling - Unable to get column medians",
f"{self.dataset_name}.{column}",
)
@_run_with_query_combiner @_run_with_query_combiner
def _get_dataset_column_stdev( def _get_dataset_column_stdev(
self, column_profile: DatasetFieldProfileClass, column: str self, column_profile: DatasetFieldProfileClass, column: str
) -> None: ) -> None:
if self.config.include_field_stddev_value: if not self.config.include_field_stddev_value:
return
try:
column_profile.stdev = str(self.dataset.get_column_stdev(column)) column_profile.stdev = str(self.dataset.get_column_stdev(column))
except Exception as e:
logger.debug(
f"Caught exception while attempting to get column stddev for column {column}. {e}"
)
self.report.report_warning(
"Profiling - Unable to get column stddev",
f"{self.dataset_name}.{column}",
)
@_run_with_query_combiner @_run_with_query_combiner
def _get_dataset_column_quantiles( def _get_dataset_column_quantiles(
self, column_profile: DatasetFieldProfileClass, column: str self, column_profile: DatasetFieldProfileClass, column: str
) -> None: ) -> None:
if self.config.include_field_quantiles: if not self.config.include_field_quantiles:
return
try:
# FIXME: Eventually we'd like to switch to using the quantile method directly. # FIXME: Eventually we'd like to switch to using the quantile method directly.
# However, that method seems to be throwing an error in some cases whereas # However, that method seems to be throwing an error in some cases whereas
# this does not. # this does not.
@ -391,6 +422,14 @@ class _SingleDatasetProfiler(BasicDatasetProfilerBase):
res["observed_value"]["values"], res["observed_value"]["values"],
) )
] ]
except Exception as e:
logger.debug(
f"Caught exception while attempting to get column quantiles for column {column}. {e}"
)
self.report.report_warning(
"Profiling - Unable to get column quantiles",
f"{self.dataset_name}.{column}",
)
@_run_with_query_combiner @_run_with_query_combiner
def _get_dataset_column_distinct_value_frequencies( def _get_dataset_column_distinct_value_frequencies(
@ -406,7 +445,9 @@ class _SingleDatasetProfiler(BasicDatasetProfilerBase):
def _get_dataset_column_histogram( def _get_dataset_column_histogram(
self, column_profile: DatasetFieldProfileClass, column: str self, column_profile: DatasetFieldProfileClass, column: str
) -> None: ) -> None:
if self.config.include_field_histogram: if not self.config.include_field_histogram:
return
try:
self.dataset.set_config_value("interactive_evaluation", True) self.dataset.set_config_value("interactive_evaluation", True)
res = self.dataset.expect_column_kl_divergence_to_be_less_than( res = self.dataset.expect_column_kl_divergence_to_be_less_than(
@ -425,6 +466,14 @@ class _SingleDatasetProfiler(BasicDatasetProfilerBase):
partition["tail_weights"][1], partition["tail_weights"][1],
], ],
) )
except Exception as e:
logger.debug(
f"Caught exception while attempting to get column histogram for column {column}. {e}"
)
self.report.report_warning(
"Profiling - Unable to get column histogram",
f"{self.dataset_name}.{column}",
)
@_run_with_query_combiner @_run_with_query_combiner
def _get_dataset_column_sample_values( def _get_dataset_column_sample_values(