ATSD Compression Test


The following test calculates the amount of disk space required to store 10+ million time:value samples in ATSD Version 17380.


Schema Compressed Total Size Sample Count Bytes per Sample
Universal Table Yes 19,590,510 10,227,570 1.9


The dataset represents 20+ years of historical minute stock trade data available from the Kibot company.

The minutely trade statistics are available for IBM stock traded on the New York Stock Exchange. The recording starts on February 1st, 1998 and lasts until the last trading day.

The data is provided in the commonly used Open,High,Low,Close,Volume (OHLCV) format.


The records can be downloaded from

The file contains over 2 million lines. The OHLC metrics contain values with up to four decimal places. The volume metric is an integer. The dates are recorded in the US/Eastern time zone.

Each row consists of 5 metrics for a given 1-minute interval:

time   = 09/08/2017 15:42
open   = 142.53
high   = 142.5399
low    = 142.49
close  = 142.49
volume = 10031


The database automatically manages schema for the inserted data.

Compression Algorithms

The storage efficiency in ATSD is a product of standard compressions algorithms such as GZIP or LZO and built-in codecs. For the purpose of this test, the compression algorithm is set to GZIP and the default ATSD codec is enabled.

Executing Tests

Download Input Data

Download the dataset.

curl --compressed -o IBM_adjusted.txt \

Verify the row count:

wc -l IBM_adjusted.txt
2045514 IBM_adjusted.txt

Launch ATSD Container

Start container with a pre-configured administrator account and port 8443 open for access.

docker run -d --name=atsd_test -p 8443:8443 \
    -e ADMIN_USER_NAME=axibase -e ADMIN_USER_PASSWORD=axibase axibase/atsd:latest

Watch the startup log until the list of open ports is displayed.

docker logs -f atsd_test
[ATSD] Waiting for ATSD to accept requests on port 8088 ... ( 5 / 60 )
[ATSD] ATSD user interface:
[ATSD] ATSD start completed. Time: 2017-09-20 22-11-07.
[ATSD] Administrator account 'axibase' created.

Download Test Files

Download CSV parser.

curl -o stock-parser.xml \

Download the test script.

curl -o \
chmod +x

Start Test

Execute the test script by specifying the ATSD container name.

Default Port

If the public port in the docker run command above is not 8443, change the ATSD_PORT variable in the script header.

./ atsd_test


Connecting to https://localhost:8443 as 'axibase'. Container: atsd_test

CSV parser imported

Uploading data file IBM_adjusted.txt


Data file uploaded IBM_adjusted.txt


Initiating data compaction
Compaction task in progress...
Compaction task in progress...
Compaction task in progress...
Compaction task in progress...
Compaction task in progress...
Compaction task in progress...
Compaction task in progress...
Compaction task in progress...
Compaction task in progress...
Compaction task complete.


Data size on disk, in bytes: 19590510


Line count in the input file: 2045514
'time:value' samples in the file: 10227570


Row count in the database: 2045514
'time:value' samples in the database: 10227570


Disk used, bytes per sample: 1.91546085727108198721
Disk used, bytes per row:    9.57730428635540993608