1937
61%
main: 61%

Ran 03 May 2019 02:56PM CUT

Jobs 4

Files 31

Run time 1min

Badge

Embed ▾

Committed 03 May 2019 02:49PM CUT coverage: 79.336% (-0.04%) from 79.376%

Build # 1937

Build Type

push

travis-ci

Committed by

manishrjain

Commit Message

Make getting key count optional in db.Tables() (#796)

Currently db.Tables() returns list of TableInfo objects. These objects
also contain keys count for the tables. Getting keys count for tables
can be expensive as it requires iterating all the tables, espcially in
large databases. db.KeySplits function(used for steaming) also uses the
same function, which does not require keys count for tables. Hence
in case of large databases, stream can take lot of time just for getting
key splits. This commit adds support for getting keys count optionally
in TableInfo objects.

I compared time for KeySplits() in with and without keys count case.
Environment:
* Ubuntu 18.10 system with 16 core, 32 GB RAM and SSD storage
* Created DB using fill command containing 500M records and 473 SSTs

I tried to create stream on DB created above (stream internally uses
db.KeySplits()). Time taken by db.KeySplits():
with keys count: 49 sec
without keys count: ~500 microsec

Run Details

5137 of 6475 relevant lines covered (79.34%)

1282327.4 hits per line

Jobs

ID	Job ID	Ran	Coverage
1	1937.1	03 May 2019 02:56PM CUT	79.06	Travis Job 1937.1
2	1937.2	03 May 2019 02:56PM CUT	79.1	Travis Job 1937.2
3	1937.3	03 May 2019 02:57PM CUT	79.1	Travis Job 1937.3
4	1937.4	03 May 2019 02:57PM CUT	79.21	Travis Job 1937.4

dgraph-io / badger / 1937
61%
main: 61%

README BADGES
x

Markdown

Textile

RDoc

HTML

Rst

Jobs

Source Files on build 1937

dgraph-io / badger / 1937 61% main: 61%

README BADGES x

Markdown

Textile

RDoc

HTML

Rst

Jobs

Source Files on build 1937

dgraph-io / badger / 1937
61%
main: 61%

README BADGES
x