c5dc48db04
feat: implement range based predicate on column
2020-11-02 11:47:14 +00:00
952959cd5f
fix: fix bug with range predicate method
2020-11-02 11:47:14 +00:00
34ba183a3e
feat: implement getting row ids from predicate
2020-11-02 11:47:14 +00:00
93dde00d04
refactor: change filter API
2020-11-02 11:47:07 +00:00
c4176a3f46
refactor: better API for row ids
2020-10-29 12:25:54 +00:00
5dd4af1286
Merge pull request #398 from influxdata/er/feat/column-encoded
...
feat: implementation of encoded value methods
2020-10-28 21:40:23 +00:00
56c56f73d6
feat: add support for all encoded values
2020-10-28 18:41:08 +00:00
e82913a186
feat: add support for encoded values
2020-10-28 17:19:01 +00:00
4a07ea8d9d
Merge pull request #397 from influxdata/er/feat/column-values
...
feat: implement methods for returning materialised values
2020-10-28 09:30:42 +00:00
1d866b448d
Merge pull request #393 from influxdata/pd-database_rules_persistence
...
feat: implement summary statistics for write buffer
2020-10-27 17:21:00 -04:00
41bc351106
chore: pr feedback fixes
2020-10-27 15:10:50 -06:00
48af64d2cb
feat: add distinct_values implementation on column
2020-10-27 18:46:41 +00:00
fef30a7cc4
feat: add values implementation on column
2020-10-27 18:46:38 +00:00
5861d6d030
refactor: make RLE values API simpler
2020-10-27 18:39:38 +00:00
ef8441054d
feat: complete signed slice conversion
2020-10-27 18:39:38 +00:00
b3d543d000
feat: extend support for converting from unsigned slices
2020-10-27 18:39:38 +00:00
d8130a6000
feat: add implementation
2020-10-27 18:39:38 +00:00
60a4e8428c
refactor: represent row ids as u32
2020-10-27 18:39:38 +00:00
6623ce9f65
feat: Add more logging for rpc, reduce low level hyper logging ( #396 )
2020-10-27 11:24:44 -04:00
cf90ebe82d
fix: Use Read instead of BufRead as no BufRead method is actually used ( #394 )
2020-10-27 07:40:27 -04:00
5105810488
feat: implement summary statistics for write buffer
...
Adds partition metadata to data types crate which also contains the summary statistics. Updates the write buffer to keep summary stats.
2020-10-26 20:03:19 -06:00
2083b41a24
refactor: rename write_buffer* to take its rightful place ( #392 )
2020-10-26 16:31:34 -04:00
1fe8f517cc
Merge pull request #389 from influxdata/pd-database_rules_persistence
...
feat: implement API for storing the server configuration
2020-10-26 15:47:48 -04:00
1e966b5153
feat: implement API for storing the server configuration in object storage
...
This adds basic API calls for persisting and loading the server configuratioon of database rules and host groups to and from object storage. It stores all the data in a single JSON file.
2020-10-26 13:43:43 -06:00
ef501871bb
feat: remove partition_store ( #387 )
2020-10-26 14:39:38 -04:00
4e1e8dbf79
chore: Upgrade version of Arrow/DataFusion (2 of 3) ( #391 )
...
* chore: Upgrade version of Arrow/DataFusion (2 of 3)
* fix: Fixup error type usage and use async stream interface
* fix: post merge fixups
2020-10-26 13:49:16 -04:00
880958d9c7
feat: switch end-to-end test to use write_buffer implementation rather than partitioned store ( #386 )
...
* feat: switch end-to-end test to use write_buffer implementation rather than partitioned store
* fix: Apply PR suggestions
2020-10-26 13:42:38 -04:00
88b9f43110
chore: Upgrade version of Arrow/DataFusion (1 of 3) ( #390 )
...
* chore: Upgrade version of Arrow/DataFusion
* fix: update code for deps
2020-10-26 11:46:02 -04:00
ac2b81e5a8
Merge pull request #385 from influxdata/er/feat/column
...
feat: initial Column stubs
2020-10-26 15:39:27 +00:00
004704c5eb
refactor: address PR feedback
2020-10-26 15:29:49 +00:00
1004854403
refactor: remove uneeded dependencies, switch to tracing from log ( #388 )
2020-10-26 06:15:47 -04:00
bee92fb33b
Merge pull request #383 from influxdata/pd-handle_replicated_write
...
feat: implement subscriptions that for subscribers to all data
2020-10-25 13:04:05 -04:00
a66bd4a738
feat: Implement measurement_fields gRPC route ( #384 )
...
* feat: Implement measurement_fields gRPC route
* fix: Apply suggestions from code review in github
Co-authored-by: Jake Goulding <jake.goulding@integer32.com>
* fix: apply code review comments locally
* fix: fix based on code review
Co-authored-by: Jake Goulding <jake.goulding@integer32.com>
2020-10-23 14:01:46 -04:00
ad7b177c26
feat: add column method stubs
2020-10-23 18:43:28 +01:00
d53a5b95a9
feat: add nullable int column
2020-10-23 15:03:02 +01:00
0e21541ba2
feat: add support for consuming arrow array
2020-10-23 15:02:38 +01:00
53dfaf6f5e
feat: create RLE columns from arrays
2020-10-23 13:41:59 +01:00
4a86f4c389
fix: ensure first rl added to column
2020-10-23 13:41:59 +01:00
2675593b35
fix: ensure dictionary only returns non-null values:
2020-10-23 13:41:59 +01:00
502773c940
feat: basic conversion of [i64] into column
2020-10-23 13:41:59 +01:00
a03f536bb7
Merge pull request #382 from influxdata/er/refactor/dict-encoding
...
feat: improve NULL support in RLE encoding
2020-10-23 13:37:45 +01:00
19911b4432
feat: implement subscriptions that for subscribers to all data
...
This implements only the most basic subscription mechanism, which basically makes it the same as replication at this point. It also simplifies subscriptions to not consider pull type. This will come back later, but will have to happen after the wal buffering is implemented.
2020-10-22 17:16:11 -04:00
45c37b22b4
Merge pull request #379 from influxdata/pd-cluster_sequence_number
...
feat: add sequence number to replicated writes
2020-10-22 14:15:02 -04:00
3914db5935
feat: add support for null/non-null rows
2020-10-22 15:30:47 +01:00
806211ed1a
refactor: re-work NULL support in RLE
2020-10-22 14:52:08 +01:00
c646f29794
feat: Add source checkout and initial build to docker CI image, don't use build cache ( #381 )
...
* feat: Add source checkout and initial build to docker CI image
* fix: remove tidbits
2020-10-22 05:50:36 -04:00
0b443fdb12
feat: add --num-threads CLI argument, allow single threaded operation ( #378 )
2020-10-21 17:12:32 -04:00
7a82fb1829
feat: add sequence number to replicated writes
...
This adds a sequence number to replicated writes. Sequence numbers are per databasea and kept only in memory, which means they'll reset whenever the process is restarted. It is up to the receivers of replicated writes to determine if they need to request any missed sequence numbers.
2020-10-21 16:11:49 -04:00
7ef17ccdaa
feat: initialise RLE with dictionary
2020-10-21 14:01:01 +01:00
daf89c7d22
Merge pull request #371 from influxdata/er/feat/dict-encoding
...
feat: add dictionary RLE encoding to Segment Store
2020-10-21 14:00:32 +01:00
Edd Robinson
Paul Dix
Paul Dix
Andrew Lamb