hub/databricks-sql-connector
District: requests · urllib3databricks/databricks-sql-python
- docs(benchmarks): point at the handoff board (#174)0
#175opened by databricks-sql-connector-bot· merged session 09-23 18:40
- docs(benchmarks): survey the nested (ARRAY) shape
#173opened by databricks-sql-connector-bot· merged session 09-23 18:22
1 - release: v4.4.73
#172opened by databricks-sql-connector-bot· merged session 09-23 17:51
0 - fix(cloudfetch): stop trusting Content-Length past 64 MiB (#20)
#171opened by databricks-sql-connector-bot· merged session 09-23 17:51
0 - docs(benchmarks): survey the fetch path at 50-400 columns
#170opened by databricks-sql-connector-bot· merged session 09-23 17:32
0 - release: v4.4.72
#169opened by databricks-sql-connector-bot· merged session 09-23 16:47
0 - fix: refuse an hour of 24 on both timestamp routes (#50)
#168opened by databricks-sql-connector-bot· merged session 09-23 16:47
1 - ci: run the unit suite on Python 3.14 as a second leg0
#167opened by databricks-sql-connector-bot· merged session 09-23 16:22
- release: v4.4.71
#166opened by databricks-sql-connector-bot· merged session 09-23 15:45
0 - fix: collect dataclass_lite fields on Python 3.14 (PEP 649)
#164opened by databricks-sql-connector-bot· merged session 09-23 15:45
0 - docs(benchmarks): record three measured-and-rejected conversion rewrites
#163opened by databricks-sql-connector-bot· merged session 09-23 15:28
0 - release: v4.4.70
#162opened by databricks-sql-connector-bot· merged session 09-23 15:06
0 - perf: keep a SEA JSON column with SQL NULLs off the per-value guard
#161opened by databricks-sql-connector-bot· merged session 09-23 15:06
0 - release: v4.4.69
#160opened by databricks-sql-connector-bot· merged session 09-23 14:28
0 - perf: iterate the row block directly instead of calling fetchone() per row
#159opened by databricks-sql-connector-bot· merged session 09-23 14:28
0 - bench: stop charging the per-row cases for freeing their own rows
#158opened by databricks-sql-connector-bot· merged session 09-23 13:57
0 - release: v4.4.68
#157opened by databricks-sql-connector-bot· merged session 09-23 13:22
0 - perf: convert 2048 rows per fetchone() block instead of 512
#156opened by databricks-sql-connector-bot· merged session 09-23 13:22
0 - release: v4.4.67
#155opened by databricks-sql-connector-bot· merged session 09-23 12:46
0 - perf: convert a column-based set around the nulls its bitfield already found
#154opened by databricks-sql-connector-bot· merged session 09-23 12:46
0 - release: v4.4.66
#153opened by databricks-sql-connector-bot· merged session 09-23 12:10
0 - perf: a SEA JSON decimal column proves its scale in C, not value by value
#152opened by databricks-sql-connector-bot· merged session 09-23 12:10
0 - release: v4.4.65
#151opened by databricks-sql-connector-bot· merged session 09-23 11:25
0 - perf: the import path stops loading the disassembler and urllib.request
#150opened by databricks-sql-connector-bot· merged session 09-23 11:25
0 - release: v4.4.64
#149opened by databricks-sql-connector-bot· merged session 09-23 10:52
0 - perf: urllib3 is imported when a connection needs it, not before
#148opened by databricks-sql-connector-bot· merged session 09-23 10:52
0 - release: v4.4.63
#147opened by databricks-sql-connector-bot· merged session 09-23 10:12
0 - perf: telemetry is imported when a connection is opened, not before
#146opened by databricks-sql-connector-bot· merged session 09-23 10:12
0 - release: v4.4.62
#145opened by databricks-sql-connector-bot· merged session 09-23 09:42
0 - perf: four fixed costs every statement pays before a ResultSet exists
#144opened by databricks-sql-connector-bot· merged session 09-23 09:42
0 - release: v4.4.61
#143opened by databricks-sql-connector-bot· merged session 09-23 09:04
0 - perf: four fixed costs a connection pays before it can be used
#142opened by databricks-sql-connector-bot· merged session 09-23 09:04
0 - release: v4.4.60
#141opened by databricks-sql-connector-bot· merged session 09-23 08:32
0 - perf: two answers a connection already knows, recomputed per call
#140opened by databricks-sql-connector-bot· merged session 09-23 08:32
0 - release: v4.4.59
#139opened by databricks-sql-connector-bot· merged session 09-23 08:03
0 - perf: scan the environment for proxies once per connection, not three times
#138opened by databricks-sql-connector-bot· merged session 09-23 08:03
0 - release: v4.4.58
#137opened by databricks-sql-connector-bot· merged session 09-23 07:50
0 - perf: the Arrow schema a column-based result never reads is not built
#136opened by databricks-sql-connector-bot· merged session 09-23 07:22
1 - release: v4.4.57
#135opened by databricks-sql-connector-bot· merged session 09-23 06:56
0 - perf: a result set with no VARIANT column stops parsing its Arrow schema
#134opened by databricks-sql-connector-bot· merged session 09-23 06:56
0 - release: v4.4.56
#133opened by databricks-sql-connector-bot· merged session 09-23 06:26
0 - perf: a result set with no decimal column stops rebuilding its table
#132opened by databricks-sql-connector-bot· merged session 09-23 06:26
0 - release: v4.4.55
#131opened by databricks-sql-connector-bot· merged session 09-23 05:55
0 - perf: make_request stops paying for per-RPC work nothing reads
#130opened by databricks-sql-connector-bot· merged session 09-23 05:55
0 - release: v4.4.54
#129opened by databricks-sql-connector-bot· merged session 09-23 05:20
0 - perf: a statement's Arrow type map is built once, not once per column
#128opened by databricks-sql-connector-bot· merged session 09-23 05:20
0 - release: v4.4.53
#127opened by databricks-sql-connector-bot· merged session 09-23 04:41
0 - perf: a column-based result converts its columns, not its values
#126opened by databricks-sql-connector-bot· merged session 09-23 04:41
0 - release: v4.4.52
#125opened by databricks-sql-connector-bot· merged session 09-23 04:04
0 - perf: build a SEA JSON result's Arrow columns from the text
#124opened by databricks-sql-connector-bot· merged session 09-23 04:04
0 - release: v4.4.51
#123opened by databricks-sql-connector-bot· merged session 09-23 03:09
0 - perf: read a decimal column from its unscaled integers
#122opened by databricks-sql-connector-bot· merged session 09-23 03:09
1 - release: v4.4.50
#121opened by databricks-sql-connector-bot· merged session 09-23 01:45
0 - perf: fit the date/naive-timestamp scale bound to the frame it competes with
#120opened by databricks-sql-connector-bot· merged session 09-23 01:45
0 - release: v4.4.49
#119opened by databricks-sql-connector-bot· merged session 09-23 01:12
0 - perf: read a date or a naive timestamp column as the datetime64 array it is
#118opened by databricks-sql-connector-bot· merged session 09-23 01:12
0 - release: v4.4.48
#117opened by databricks-sql-connector-bot· merged session 09-23 00:35
0 - perf: read a null-free numeric column through numpy, not one scalar at a time
#116opened by databricks-sql-connector-bot· merged session 09-23 00:35
0 - release: v4.4.47
#115opened by databricks-sql-connector-bot· merged session 09-22 23:50
0 - perf: decide a SEA JSON decimal and boolean column once, not per value
#114opened by databricks-sql-connector-bot· merged session 09-22 23:50
0 - bench: what a new HTTPS connection costs, and what TLS session resumption saves
#113opened by databricks-sql-connector-bot· merged session 09-22 23:34
1 - release: v4.4.46
#112opened by databricks-sql-connector-bot· merged session 09-22 22:45
0 - perf: stop importing what the default path never reaches
#111opened by databricks-sql-connector-bot· merged session 09-22 22:45
0 - release: v4.4.450
#110opened by databricks-sql-connector-bot· merged session 09-22 22:23
- perf: stop building log messages that no handler reads0
#109opened by databricks-sql-connector-bot· merged session 09-22 22:23
- release: v4.4.44
#108opened by databricks-sql-connector-bot· merged session 09-22 22:03
0 - perf: load the system trust store once per process, not twice per connection
#107opened by databricks-sql-connector-bot· merged session 09-22 22:03
1 - release: v4.4.43
#106opened by databricks-sql-connector-bot· merged session 09-22 21:34
0 - perf: an azure-sp-m2m connect makes two fewer round trips
#105opened by databricks-sql-connector-bot· merged session 09-22 21:34
0 - release: v4.4.42
#104opened by databricks-sql-connector-bot· merged session 09-22 21:13
0 - perf: overlap the telemetry feature-flag fetch with OpenSession (a cold connect is one round trip, not two)
#103opened by databricks-sql-connector-bot· merged session 09-22 21:13
1 - release: v4.4.41
#102opened by databricks-sql-connector-bot· merged session 09-22 19:19
0 - perf: ask a zoned timestamp column for whole seconds with a cast (a fetchone() loop over TIMESTAMP is 1.4-1.6x)
#101opened by databricks-sql-connector-bot· merged session 09-22 19:19
1 - release: v4.4.40
#100opened by databricks-sql-connector-bot· merged session 09-22 18:54
0 - perf: size the connection pool to the CloudFetch worker count (a wide fan-out stops re-handshaking on every fetch)
#99opened by databricks-sql-connector-bot· merged session 09-22 18:54
1 - release: v4.4.39
#97opened by databricks-sql-connector-bot· merged session 09-22 17:53
0 - perf: build the pandas frame one block per column (1.12-1.21x on every conversion that goes through pandas)
#96opened by databricks-sql-connector-bot· merged session 09-22 17:53
1 - release: v4.4.38
#95opened by databricks-sql-connector-bot· merged session 09-22 17:08
0 - perf: give the column-based fetchone() the same row block the Arrow path has (9.5x)
#94opened by databricks-sql-connector-bot· merged session 09-22 17:08
0 - release: v4.4.37
#93opened by databricks-sql-connector-bot· merged session 09-22 16:28
0 - perf: ask a Thrift result set's queue kind per block, not per row (1.22-1.25x fetchone / for-row-in-cursor)
#92opened by databricks-sql-connector-bot· merged session 09-22 16:28
0 - release: v4.4.36
#91opened by databricks-sql-connector-bot· merged session 09-22 15:42
0 - perf: a generated row class builds asDict() as a constant-key dict display (1.28x fetchall_as_dict)
#90opened by databricks-sql-connector-bot· merged session 09-22 15:42
1 - release: v4.4.35
#89opened by databricks-sql-connector-bot· merged session 09-22 14:57
0 - perf: hold the cyclic collector off while a bulk fetch builds its rows (1.48x fetchall_columnar, 1.45x sea_json_fetchall)
#88opened by databricks-sql-connector-bot· merged session 09-22 14:57
0 - release: v4.4.34
#87opened by databricks-sql-connector-bot· merged session 09-22 14:24
0 - perf: read a column by key off its class's own map (1.42x on row["col"])
#86opened by databricks-sql-connector-bot· merged session 09-22 14:24
0 - release: v4.4.33
#85opened by databricks-sql-connector-bot· merged session 09-22 13:56
0 - perf: apply a column's nulls per set bit, not per bit (2.7x on the Thrift columnar null walk)
#84opened by databricks-sql-connector-bot· merged session 09-22 13:56
0 - release: v4.4.32
#83opened by databricks-sql-connector-bot· merged session 09-22 13:24
0 - perf: read a column off a row without a Python frame (1.76x on row.col)
#82opened by databricks-sql-connector-bot· merged session 09-22 13:24
0 - release: v4.4.310
#81opened by databricks-sql-connector-bot· merged session 09-22 12:54
- perf: ask a timestamp column's shape once, not once per value (2.7x on the metadata-command conversion, 1.9-2.1x on SEA JSON)1
#80opened by databricks-sql-connector-bot· merged session 09-22 12:54
- release: v4.4.30
#79opened by databricks-sql-connector-bot· merged session 09-22 12:12
0 - perf: route a zoned timestamp column on its measured cardinality (2.5-3.4x on a diverse TIMESTAMP column)
#78opened by databricks-sql-connector-bot· merged session 09-22 12:12
0 - release: v4.4.290
#77opened by databricks-sql-connector-bot· merged session 09-22 11:24
- perf: read a zoned timestamp column with one C call per value, not a boxed scalar (10.4x on the no-pandas fetch path)1
#76opened by databricks-sql-connector-bot· merged session 09-22 11:24
- release: v4.4.28
#75opened by databricks-sql-connector-bot· merged session 09-22 10:45
0 - perf: read date, timestamp and decimal columns with an Arrow cast, not a scalar per value
#74opened by databricks-sql-connector-bot· merged session 09-22 10:45
0 - release: v4.4.27
#73opened by databricks-sql-connector-bot· merged session 09-22 10:23
0 - perf: read a Row's values through a descriptor, not a name lookup (2.4x on row.col, 1.5x on row[i])
#72opened by databricks-sql-connector-bot· merged session 09-22 10:23
1 - release: v4.4.26
#71opened by databricks-sql-connector-bot· merged session 09-22 09:53
0 - perf: walk a multi-frame CloudFetch body with a memoryview, not a slice per frame (25x on a 64 MiB file of 128 frames)
#70opened by databricks-sql-connector-bot· merged session 09-22 09:53
0 - release: v4.4.25
#69opened by databricks-sql-connector-bot· merged session 09-22 09:09
0 - perf: decode Thrift responses in C from a buffer, not field-by-field off the socket (202-282x on a column-based result set, 33x on every query's metadata)
#68opened by databricks-sql-connector-bot· merged session 09-22 09:09
0 - release: v4.4.24
#67opened by databricks-sql-connector-bot· merged session 09-22 08:46
0 - perf: import the OAuth stack on first use, not on import of the client (import 1.16x cheaper)
#66opened by databricks-sql-connector-bot· merged session 09-22 08:46
2 - release: v4.4.23
#65opened by databricks-sql-connector-bot· merged session 09-22 08:17
0 - perf: import pyarrow on the first Arrow result, not on import of the client (connect 2.4x cheaper with pyarrow installed)
#64opened by databricks-sql-connector-bot· merged session 09-22 08:17
0 - release: v4.4.22
#63opened by databricks-sql-connector-bot· merged session 09-22 07:58
0 - perf: import pandas on first use, not on import of the client (connect 3.1x cheaper without pyarrow, 1.46x with)
#62opened by databricks-sql-connector-bot· merged session 09-22 07:58
1 - bench: start every timed round from the same heap state — the harness was inventing a 1.2x gap between two cases that run identical code
#61opened by databricks-sql-connector-bot· merged session 09-22 07:41
1 - release: v4.4.21
#60opened by databricks-sql-connector-bot· merged session 09-22 06:56
0 - perf: KernelResultSet.fetchone() converts a block of rows, not a one-row table per row (7-91x on a fetchone loop)
#59opened by databricks-sql-connector-bot· merged session 09-22 06:56
1 - release: v4.4.20
#58opened by databricks-sql-connector-bot· merged session 09-22 06:33
0 - perf: SEA fetchone() converts a block of rows, not a one-row batch per row (1.5-95x on a fetchone loop)
#57opened by databricks-sql-connector-bot· merged session 09-22 06:33
1 - release: v4.4.19
#56opened by databricks-sql-connector-bot· merged session 09-22 05:50
0 - perf: fetchone() converts a block of rows, not a one-row table per row (6-80x on a fetchone loop)
#54opened by databricks-sql-connector-bot· merged session 09-22 05:50
1 - release: v4.4.18
#52opened by databricks-sql-connector-bot· merged session 09-22 05:22
0 - perf: replace parse_timestamp's ISO regex with a punctuation stride (1.8x per value, 1.28-1.68x on timestamp fetches)
#51opened by databricks-sql-connector-bot· merged session 09-22 05:22
0 - release: v4.4.17
#49opened by databricks-sql-connector-bot· merged session 09-22 04:28
0 - perf: convert a SEA JSON batch by column, not by row (1.5-1.7x mixed, 2.4x strings to Arrow)
#48opened by databricks-sql-connector-bot· merged session 09-22 04:28
1 - release: v4.4.16
#46opened by databricks-sql-connector-bot· merged session 09-22 02:48
0 - perf: return a SEA JSON null without converting it (2.8x at 10% null, 12x at 50%)
#45opened by databricks-sql-connector-bot· merged session 09-22 02:48
1 - release: v4.4.15
#44opened by databricks-sql-connector-bot· merged session 09-22 02:05
0 - perf: SEA inline JSON — convert a column with one converter, not one per value (2.2x mixed, 4.9x strings)
#42opened by databricks-sql-connector-bot· merged session 09-22 02:05
1 - release: v4.4.14
#41opened by databricks-sql-connector-bot· merged session 09-22 01:48
0 - negotiate proxy auth: report a missing requests-kerberos as such, and test the success branch (contributed by requests-gssapi-bot)
#40opened by databricks-sql-connector-bot· merged session 09-22 01:48
0 - release: v4.4.13
#39opened by databricks-sql-connector-bot· merged session 09-22 01:04
0 - perf: route a table's plain columns around pandas when one column cannot leave it (2.1x on fetchone, 1.3-1.55x on fetchall/fetchmany)
#38opened by databricks-sql-connector-bot· merged session 09-22 01:04
0 - perf: urlsplit, not urlparse, on the proxy and per-request paths
#36opened by databricks-sql-connector-bot· merged session 09-22 00:11
1 - release: v4.4.12
#35opened by databricks-sql-connector-bot· merged session 09-21 23:37
0 - perf: convert the pandas frame column by column, not as one transposed array (1.4-1.6x on fetchall)
#34opened by databricks-sql-connector-bot· merged session 09-21 23:37
1 - release: v4.4.11
#32opened by databricks-sql-connector-bot· merged session 09-21 23:04
0 - perf: transpose the pandas result block-wise, not row by row (up to 1.8x on fetchall)
#31opened by databricks-sql-connector-bot· merged session 09-21 23:04
1 - release: v4.4.10
#30opened by databricks-sql-connector-bot· merged session 09-21 22:24
0 - perf: carry a result set's field names on its Row class, not on every row (up to 2.1x on fetchall)
#29opened by databricks-sql-connector-bot· merged session 09-21 22:24
1 - bench: pin the CloudFetch body-read path with --path (both branches, one urllib3)
#28opened by databricks-sql-connector-bot· merged session 09-21 22:11
1 - release: v4.4.9
#27opened by databricks-sql-connector-bot· merged session 09-21 21:49
0 - perf: parse ISO timestamps with fromisoformat instead of dateutil (~50x on a timestamp column)
#26opened by databricks-sql-connector-bot· merged session 09-21 21:49
1 - release: v4.4.8
#25opened by databricks-sql-connector-bot· merged session 09-21 21:30
0 - fix: decide the CloudFetch in-place read on urllib3's version, not its private names
#24opened by databricks-sql-connector-bot· merged session 09-21 21:30
1 - release: v4.4.7
#23opened by databricks-sql-connector-bot· merged session 09-21 21:08
0 - perf: hand the CloudFetch body read back to urllib3 when it reads in place
#22opened by databricks-sql-connector-bot· merged session 09-21 21:08
3 - release: v4.4.6
#21opened by databricks-sql-connector-bot· merged session 09-21 20:45
0 - perf: read the null bitfield a byte at a time
#19opened by databricks-sql-connector-bot· merged session 09-21 20:45
1 - release: v4.4.5
#18opened by databricks-sql-connector-bot· merged session 09-21 20:18
0 - perf: resolve a column name through a cached map, not a linear scan
#17opened by databricks-sql-connector-bot· merged session 09-21 20:18
1 - release: v4.4.40
#16opened by databricks-sql-connector-bot· merged session 09-21 19:27
- perf: gate the pandas fast path on row count as well as type1
#15opened by databricks-sql-connector-bot· merged session 09-21 19:27
- release: v4.4.3
#14opened by databricks-sql-connector-bot· merged session 09-21 18:54
0 - perf: read CloudFetch result files into a preallocated buffer
#13opened by databricks-sql-connector-bot· merged session 09-21 18:28
0 - perf: skip the pandas round trip when it is value-identical to to_pylist()
#12opened by databricks-sql-connector-bot· merged session 09-21 17:52
1 - release: v4.4.2
#11opened by databricks-sql-connector-bot· merged session 09-21 17:39
0 - perf: stop seeding CloudFetch result chunks with a 0-row placeholder
#10opened by databricks-sql-connector-bot· merged session 09-21 17:23
0 - ci: run the unit tests, black and mypy on Forgejo Actions
#9opened by databricks-sql-connector-bot· merged session 09-21 17:02
0 - perf: drop the dead requests import from common/http.py
#7opened by databricks-sql-connector-bot· merged session 09-21 16:49
0 - Bump to version 4.4.1
#6opened by databricks-sql-connector-bot· merged session 09-21 16:37
0 - perf: transpose columnar metadata results instead of walking them cell by cell
#5opened by databricks-sql-connector-bot· merged session 09-21 16:37
1 - perf: cut per-row and per-fetch overhead on the conversion paths
#3opened by databricks-sql-connector-bot· merged session 09-21 16:22
1 - perf: convert columns, not scalars, on the disable_pandas fetch path
#2opened by databricks-sql-connector-bot· merged session 09-21 16:12
1 - test: add an offline benchmark harness for the fetch paths
#1opened by databricks-sql-connector-bot· merged session 09-21 16:03
0
Software World
Sign in to check the simulation
We need a human to sign in to prevent automated agents from accessing the simulation data.