Changelog¶
1.25.0 (2020-06-06)¶
Features¶
add BigQuery storage client support to DB API (#36) (ba9b2f8)
bigquery: add support of model for extract job (#71) (4a7a514)
add HOUR support for time partitioning interval (#91) (0dd90b9)
bigquery: expose start index parameter for query result (#121) (be86de3)
bigquery: unit and system test for dataframe with int column with Nan values (#39) (5fd840e)
1.24.0¶
02-03-2020 01:38 PST
Implementation Changes¶
New Features¶
Check
rows
argument type ininsert_rows()
. (#10174)Check
json_rows
arg type ininsert_rows_json()
. (#10162)Make
RowIterator.to_dataframe_iterable()
method public. (#10017)Add retry parameter to public methods where missing. (#10026)
Add timeout parameter to Client and Job public methods. (#10002)
Add timeout parameter to
QueryJob.done()
method. (#9875)Add
create_bqstorage_client
parameter toto_dataframe()
andto_arrow()
methods. (#9573)
Dependencies¶
Fix minimum versions of
google-cloud-core
andgoogle-resumable-media
dependencies. (#10016)
Documentation¶
1.22.0¶
11-13-2019 12:23 PST
Implementation Changes¶
New Features¶
Documentation¶
Fix link anchors in external config docstrings. (#9627)
Add python 2 sunset banner to documentation. (#9036)
Add table create sample using integer range partitioning. (#9478)
Document how to achieve higher write limit and add tests. (#9574)
Add code sample for scripting. (#9537)
Rewrite docs in Google style, part 2. (#9481)
Use multi-regional key path for CMEK in snippets. (#9523)
1.21.0¶
10-16-2019 10:33 PDT
New Features¶
add ability to pass in a table ID instead of a query to the
%%bigquery
magic (#9170)add support for custom
QueryJobConfig
inBigQuery.cursor.execute
method (#9278)store
QueryJob
to destination var on error in%%bigquery
magic (#9245)add script statistics to job resource (#9428)
add support for sheets ranges (#9416)
add support for listing jobs by parent job (#9225)
expose customer managed encryption key for ML models (#9302)
add
Dataset.default_partition_expiration_ms
andTable.require_partition_filter
properties (#9464)
Documentation¶
1.20.0¶
09-13-2019 11:22 PDT
Implementation Changes¶
New Features¶
Add
Model.encryption_config
. (via synth) (#9214)Add
Client.insert_rows_from_dataframe()
method (#9162)Add support for array parameters to
Cursor.execute()
. (#9189)Add support for project IDs with org prefix to
Table.from_string()
factory. (#9161)Add
--max_results
option to Jupyter magics (#9169)Autofetch table schema on load if not provided. (#9108)
Add
max_results
parameter toQueryJob.result()
. (#9167)
1.19.0¶
09-03-2019 14:33 PDT
Implementation Changes¶
Raise when unexpected fields are present in the
LoadJobConfig.schema
when callingload_table_from_dataframe
. (#9096)Determine the schema in
load_table_from_dataframe
based on dtypes. (#9049)Raise helpful error when loading table from dataframe with
STRUCT
columns. (#9053)Fix schema recognition of struct field types. (#9001)
Fix deserializing
None
inQueryJob
for queries with parameters. (#9029)
New Features¶
Include indexes in table written by
load_table_from_dataframe
, only if fields corresponding to indexes are present inLoadJobConfig.schema
. (#9084)Add
client_options
to constructor. (#8999)Add
--dry_run
option to%%bigquery
magic. (#9067)Add
load_table_from_json()
method to create a table from a list of dictionaries. (#9076)Allow subset of schema to be passed into
load_table_from_dataframe
. (#9064)Add support for unsetting
LoadJobConfig.schema
. (#9077)Add support to
Dataset
for project IDs containing an org prefix. (#8877)Add enum with SQL type names allowed to be used in
SchemaField
. (#9040)
Documentation¶
1.18.0¶
08-08-2019 12:28 PDT
New Features¶
Add
bqstorage_client
param toQueryJob.to_arrow()
(#8693)Include SQL query and job ID in exception messages. (#8748)
Allow using TableListItem to construct a Table object. (#8738)
Add StandardSqlDataTypes enum to BigQuery (#8782)
Add
to_standard_sql()
method to SchemaField (#8880)Add debug logging statements to track when BQ Storage API is used. (#8838)
Hide error traceback in BigQuery cell magic (#8808)
Allow choice of compression when loading from dataframe (#8938)
Additional clustering metrics for BQML K-means models (via synth). (#8945)
Documentation¶
Internal / Testing Changes¶
Fix several pytest “skip if” markers (#8694)
Update tests to support conversion of NaN as NULL in pyarrow
0.14.*
. (#8785)Mock external calls in one of BigQuery unit tests (#8727)
Set IPython user agent when running queries with IPython cell magic (#8713)
Use configurable bucket name for GCS samples data in systems tests. (#8783)
Move
maybe_fail_import()
to top level test utils (#8840)Set BQ Storage client user-agent when in Jupyter cell (#8734)
1.15.0¶
06-14-2019 10:10 PDT
1.14.0¶
06-04-2019 11:11 PDT
1.13.0¶
05-31-2019 10:22 PDT
1.12.1¶
05-21-2019 11:16 PDT
1.12.0¶
05-16-2019 11:25 PDT
Implementation Changes¶
Remove duplicates from index on pandas DataFrames returned by
to_dataframe()
. (#7953)Prevent error when time partitioning is populated with empty dict (#7904)
Preserve order in
to_dataframe
with BQ Storage from queries containingORDER BY
(#7793)Respect
progress_bar_type
into_dataframe
when used with BQ Storage API (#7697)Refactor QueryJob.query to read from resource dictionary (#7763)
Close the
to_dataframe
progress bar when finished. (#7757)Ensure that
KeyboardInterrupt
duringto_dataframe
no longer hangs. (#7698)Raise ValueError when BQ Storage is required but missing (#7726)
Make
total_rows
available on RowIterator before iteration (#7622)Avoid masking auth errors in
to_dataframe
with BQ Storage API (#7674)
New Features¶
Dependencies¶
1.11.0¶
04-03-2019 19:33 PDT
New Features¶
Enable fastparquet support by using temporary file in
load_table_from_dataframe
(#7545)Allow string for copy sources, query destination, and default dataset (#7560)
Add
progress_bar_type
argument toto_dataframe
to usetqdm
to display a progress bar (#7552)Call
get_table
inlist_rows
if the schema is not available (#7621)Fallback to BQ API when there are problems reading from BQ Storage. (#7633)
Add methods for Models API (#7562)
Add option to use BigQuery Storage API from IPython magics (#7640)
Documentation¶
1.9.0¶
02-04-2019 13:28 PST
1.8.1¶
12-17-2018 17:53 PST
1.8.0¶
12-10-2018 12:39 PST
Implementation Changes¶
Add option to use BQ Storage API with
to_dataframe
(#6854)Fix exception type in comment (#6847)
Add
to_bqstorage
to convert from Table[Reference] google-cloud-bigquery-storage reference (#6840)Import
iam.policy
fromgoogle.api_core
. (#6741)Add avro logical type control for load jobs. (#6827)
Allow setting partition expiration to ‘None’. (#6823)
Add
retry
argument to_AsyncJob.result
. (#6302)
1.6.0¶
1.5.1¶
Implementation Changes¶
Retry ‘502 Bad Gateway’ errors by default. (#5930)
Avoid pulling entire result set into memory when constructing dataframe. (#5870)
Add support for retrying unstructured 429 / 500 / 502 responses. (#6011)
Populate the jobReference from the API response. (#6044)
Documentation¶
Prepare documentation for repo split (#5955)
Fix leakage of bigquery/spanner sections into sidebar menu. (#5986)
Internal / Testing Changes¶
Test pandas support under Python 3.7. (#5857)
Nox: use inplace installs (#5865)
Update system test to use test data in bigquery-public-data. (#5965)
1.5.0¶
Implementation Changes¶
Make ‘Table.location’ read-only. (#5687)
New Features¶
Add ‘clustering_fields’ properties. (#5630)
Add support for job labels (#5654)
Add ‘QueryJob.estimated_bytes_processed’ property (#5655)
Add support/tests for loading tables from ‘gzip.GzipFile’. (#5711)
Add ‘ExternalSourceFormat’ enum. (#5674)
Add default location to client (#5678)
Documentation¶
Fix typo in CopyJob sources docstring (#5690)
Internal / Testing Changes¶
Add/refactor snippets for managing BigQuery jobs (#5631)
Reenable systests for ‘dataset.update’/’table.update’. (#5732)
1.4.0¶
Implementation Changes¶
Add ‘internalError’ to retryable error reasons. (#5599)
Don’t raise exception if viewing CREATE VIEW DDL results (#5602)
New Features¶
Add Orc source format support and samples (#5500)
Move ‘DEFAULT_RETRY’ (w/ its predicate) to a new public ‘retry’ module. (#5552)
Allow listing rows on an empty table. (#5584)
Documentation¶
Add load_table_from_dataframe() to usage docs and changelog and dedents snippets in usage page (#5501)
Add samples for query external data sources (GCS & Sheets) (#5491)
Add BigQuery authorized view samples (#5515)
Update docs to show pyarrow as the only dependency of load_table_from_dataframe() (#5582)
Internal / Testing Changes¶
Add missing explict coverage for ‘_helpers’ (#5550)
Skip update_table and update_dataset tests until etag issue is resolved. (#5590)
1.3.0¶
New Features¶
NUMERIC type support (#5331)
Add timeline and top-level slot-millis to query statistics. (#5312)
Add additional statistics to query plan stages. (#5307)
Add
client.load_table_from_dataframe()
(#5387)
Documentation¶
Use autosummary to split up API reference docs (#5340)
Fix typo in Client docstrings (#5342)
Internal / Testing Changes¶
Prune systests identified as reduntant to snippets. (#5365)
Modify system tests to use prerelease versions of grpcio (#5304)
Improve system test performance (#5319)
1.2.0¶
Implementation Changes¶
Switch
list_partitions
helper to a direct metatable read (#5273)Fix typo in
Encoding.ISO_8859_1
enum value (#5211)
New Features¶
Add UnknownJob type for redacted jobs. (#5281)
Add project parameter to
list_datasets
andlist_jobs
(#5217)Add from_string factory methods to Dataset and Table (#5255)
Add column based time partitioning (#5267)
Documentation¶
Standardize docstrings for constants (#5289)
Fix docstring / impl of
ExtractJob.destination_uri_file_counts
. (#5245)
Internal / Testing Changes¶
Add testing support for Python 3.7; remove testing support for Python 3.4. (#5295)
1.1.0¶
New Features¶
Add
client.get_service_account_email
(#5203)
Documentation¶
Update samples and standardize region tags (#5195)
Internal / Testing Changes¶
Fix trove classifier to be Production/Stable
Don’t suppress ‘dots’ output on test (#5202)
0.32.0¶
:warning: Interface changes¶
Use
job.configuration
resource for XXXJobConfig classes (#5036)
Interface additions¶
Add
page_size
parameter forlist_rows
and use in DB-API forarraysize
(#4931)Add IPython magics for running queries (#4983)
Documentation¶
Add job string constant parameters in init and snippets documentation (#4987)
Internal / Testing changes¶
Specify IPython version 5.5 when running Python 2.7 tests (#5145)
Move all Dataset property conversion logic into properties (#5130)
Remove unnecessary _Table class from test_job.py (#5126)
Use explicit bytes to initialize ‘BytesIO’. (#5116)
Make SchemaField be able to include description via from_api_repr method (#5114)
Remove _ApiResourceProperty class (#5107)
Add dev version for 0.32.0 release (#5105)
StringIO to BytesIO (#5101)
Shorten snippets test name (#5091)
Don’t use
selected_fields
for listing query result rows (#5072)Add location property to job classes. (#5071)
Use autospec for Connection in tests. (#5066)
Add Parquet SourceFormat and samples (#5057)
Remove test_load_table_from_uri_w_autodetect_schema_then_get_job because of duplicate test in snippets (#5004)
Fix encoding variable and strings UTF-8 and ISO-8859-1 difference documentation (#4990)
0.31.0¶
Interface additions¶
Add support for
EncryptionConfiguration
(#4845)
Implementation changes¶
Allow listing/getting jobs even when there is an “invalid” job. (#4786)
Dependencies¶
The minimum version for
google-api-core
has been updated to version 1.0.0. This may cause some incompatibility with older google-cloud libraries, you will need to update those libraries if you have a dependency conflict. (#4944, #4946)
Documentation¶
Update format in
Table.full_table_id
andTableListItem.full_table_id
docstrings. (#4906)
Testing and internal changes¶
Install local dependencies when running lint (#4936)
Re-enable lint for tests, remove usage of pylint (#4921)
Normalize all setup.py files (#4909)
Remove unnecessary debug print from tests (#4907)
Use constant strings for job properties in tests (#4833)
0.30.0¶
This is the release candidate for v1.0.0.
Interface changes / additions¶
Add
delete_contents
todelete_dataset
. (#4724)
Bugfixes¶
Add handling of missing properties in
SchemaField.from_api_repr()
. (#4754)Fix missing return value in
LoadJobConfig.from_api_repr
. (#4727)
Documentation¶
Minor documentation and typo fixes. (#4782, #4718, #4784, #4835, #4836)
0.29.0¶
Interface changes / additions¶
Add
to_dataframe()
method to row iterators. When Pandas is installed this method returns aDataFrame
containing the query’s or table’s rows. (#4354)Iterate over a
QueryJob
to wait for and get the query results. (#4350)Add
Table.reference
andDataset.reference
properties to get theTableReference
orDatasetReference
corresponding to thatTable
orDataset
, respectively. (#4405)Add
Row.keys()
,Row.items()
, andRow.get()
. This makesRow
act more like a built-in dictionary. (#4393, #4413)
Interface changes / breaking changes¶
Add
Client.insert_rows()
andClient.insert_rows_json()
, deprecateClient.create_rows()
andClient.create_rows_json()
. (#4657)Add
Client.list_tables
, deprecateClient.list_dataset_tables
. (#4653)Client.list_tables
returns an iterators ofTableListItem
. The API only returns a subset of properties of a table when listing. (#4427)Remove
QueryJob.query_results()
. UseQueryJob.result()
instead. (#4652)Remove
Client.query_rows()
. UseClient.query()
instead. (#4429)Client.list_datasets
returns an iterator ofDatasetListItem
. The API only returns a subset of properties of a dataset when listing. (#4439)
0.28.0¶
0.28.0 significantly changes the interface for this package. For examples of the differences between 0.28.0 and previous versions, see Migrating to the BigQuery Python client library 0.28. These changes can be summarized as follows:
Query and view operations default to the standard SQL dialect. (#4192)
Client functions related to jobs, like running queries, immediately start the job.
Functions to create, get, update, delete datasets and tables moved to the client class.
Fixes¶
Populate timeout parameter correctly for queries (#4209)
Automatically retry idempotent RPCs (#4148, #4178)
Parse timestamps in query parameters using canonical format (#3945)
Parse array parameters that contain a struct type. (#4040)
Support Sub Second Datetimes in row data (#3901, #3915, #3926), h/t @page1
Interface changes / additions¶
Support external table configuration (#4182) in query jobs (#4191) and tables (#4193).
New
Row
class allows for access by integer index like a tuple, string index like a dictionary, or attribute access like an object. (#4149)Add option for job ID generation with user-supplied prefix (#4198)
Add support for update of dataset access entries (#4197)
Add support for atomic read-modify-write of a dataset using etag (#4052)
Add support for labels to
Dataset
(#4026)Add support for labels to
Table
(#4207)Add
Table.streaming_buffer
property (#4161)Add
TableReference
class (#3942)Add
DatasetReference
class (#3938, #3942, #3993)Add
ExtractJob.destination_uri_file_counts
property. (#3803)Add
client.create_rows_json()
to bypass conversions on streaming writes. (#4189)Add
client.get_job()
to get arbitrary jobs. (#3804, #4213)Add filter to
client.list_datasets()
(#4205)Add
QueryJob.undeclared_query_parameters
property. (#3802)Add
QueryJob.referenced_tables
property. (#3801)Add new scalar statistics properties to
QueryJob
(#3800)Add
QueryJob.query_plan
property. (#3799)
Interface changes / breaking changes¶
Remove
client.run_async_query()
, useclient.query()
instead. (#4130)Remove
client.run_sync_query()
, useclient.query_rows()
instead. (#4065, #4248)Make
QueryResults
read-only. (#4094, #4144)Make
get_query_results
private. Return rows forQueryJob.result()
(#3883)Move
*QueryParameter
andUDFResource
classes toquery
module (also exposed inbigquery
module). (#4156)
Changes to tables¶
Remove
client
fromTable
class (#4159)Remove
table.exists()
(#4145)Move
table.list_parations
toclient.list_partitions
(#4146)Move
table.upload_from_file
toclient.load_table_from_file
(#4136)Move
table.update()
andtable.patch()
toclient.update_table()
(#4076)Move
table.insert_data()
toclient.create_rows()
. Automatically generates row IDs if not supplied. (#4151, #4173)Move
table.fetch_data()
toclient.list_rows()
(#4119, #4143)Move
table.delete()
toclient.delete_table()
(#4066)Move
table.create()
toclient.create_table()
(#4038, #4043)Move
table.reload()
toclient.get_table()
(#4004)Rename
Table.name
attribute toTable.table_id
(#3959)Table
constructor takes aTableReference
as parameter (#3997)
Changes to datasets¶
Remove
client
fromDataset
class (#4018)Remove
dataset.exists()
(#3996)Move
dataset.list_tables()
toclient.list_dataset_tables()
(#4013)Move
dataset.delete()
toclient.delete_dataset()
(#4012)Move
dataset.patch()
anddataset.update()
toclient.update_dataset()
(#4003)Move
dataset.create()
toclient.create_dataset()
(#3982)Move
dataset.reload()
toclient.get_dataset()
(#3973)Rename
Dataset.name
attribute toDataset.dataset_id
(#3955)client.dataset()
returns aDatasetReference
instead ofDataset
. (#3944)Rename class:
dataset.AccessGrant -> dataset.AccessEntry
. (#3798)dataset.table()
returns aTableReference
instead of aTable
(#4014)Dataset
constructor takes a DatasetReference (#4036)
Changes to jobs¶
Make
job.begin()
method private. (#4242)Add
LoadJobConfig
class and modifyLoadJob
(#4103, #4137)Add
CopyJobConfig
class and modifyCopyJob
(#4051, #4059)Type of Job’s and Query’s
default_dataset
changed fromDataset
toDatasetReference
(#4037)Rename
client.load_table_from_storage()
toclient.load_table_from_uri()
(#4235)Rename
client.extract_table_to_storage
toclient.extract_table()
. Method starts the extract job immediately. (#3991, #4177)Rename
XJob.name
toXJob.job_id
. (#3962)Rename job classes.
LoadTableFromStorageJob -> LoadJob
andExtractTableToStorageJob -> jobs.ExtractJob
(#3797)
Dependencies¶
Updating to
google-cloud-core ~= 0.28
, in particular, thegoogle-api-core
package has been moved out ofgoogle-cloud-core
. (#4221)
PyPI: https://pypi.org/project/google-cloud-bigquery/0.28.0/
0.27.0¶
Remove client-side enum validation. (#3735)
Add
Table.row_from_mapping
helper. (#3425)Move
google.cloud.future
togoogle.api.core
(#3764)Fix
__eq__
and__ne__
. (#3765)Move
google.cloud.iterator
togoogle.api.core.page_iterator
(#3770)nullMarker
support for BigQuery Load Jobs (#3777), h/t @leondealmeidaAllow
job_id
to be explicitly specified in DB-API. (#3779)Add support for a custom null marker. (#3776)
Add
SchemaField
serialization and deserialization. (#3786)Add
get_query_results
method to the client. (#3838)Poll for query completion via
getQueryResults
method. (#3844)Allow fetching more than the first page when
max_results
is set. (#3845)
PyPI: https://pypi.org/project/google-cloud-bigquery/0.27.0/
0.26.0¶
Notable implementation changes¶
Using the
requests
transport attached to a Client for for resumable media (i.e. downloads and uploads) (#3705) (this relates to thehttplib2
torequests
switch)
Interface changes / additions¶
Adding
autodetect
property onLoadTableFromStorageJob
to enable schema autodetection. (#3648)Implementing the Python Futures interface for Jobs. Call
job.result()
to wait for jobs to complete instead of polling manually on the job status. (#3626)Adding
is_nullable
property onSchemaField
. Can be used to check if a column is nullable. (#3620)job_name
argument added toTable.upload_from_file
for setting the job ID. (#3605)Adding
google.cloud.bigquery.dbapi
package, which implements PEP-249 DB-API specification. (#2921)Adding
Table.view_use_legacy_sql
property. Can be used to create views with legacy or standard SQL. (#3514)
Interface changes / breaking changes¶
Removing
results()
method from theQueryJob
class. Usequery_results()
instead. (#3661)SchemaField
is now immutable. It is also hashable so that it can be used in sets. (#3601)
Dependencies¶
Updating to
google-cloud-core ~= 0.26
, in particular, the underlying HTTP transport switched fromhttplib2
torequests
(#3654, #3674)Adding dependency on
google-resumable-media
for loading BigQuery tables from local files. (#3555)
Packaging¶
Fix inclusion of
tests
(vs.unit_tests
) inMANIFEST.in
(#3552)Updating
author_email
insetup.py
togoogleapis-publisher@google.com
. (#3598)
PyPI: https://pypi.org/project/google-cloud-bigquery/0.26.0/