Skip to content

‣ DSBulk

📖 Reference Documentations and resources
  1. 📖 DSBulks Docs - Reference documentation
  2. 📖 DataStax Docs - Reference Documentation

A - Overview

📘 What is DSBulk ?

The DataStax Bulk Loader tool (DSBulk) is a unified tool for loading into and unloading from Cassandra-compatible storage engines, such as OSS Apache Cassandra®, DataStax Astra and DataStax Enterprise (DSE).

Out of the box, DSBulk provides the ability to:

  • Load (import) large amounts of data into the database efficiently and reliably;
  • Unload (export) large amounts of data from the database efficiently and reliably;
  • Count elements in a database table: how many rows in total, how many rows per replica and per token range, and how many rows in the top N largest partitions.
# Load data
dsbulk load <options>

# Unload data
dsbulk unload <options>

# Count rows
dsbulk count <options>

Currently, CSV and Json formats are supported for both loading and unloading data.

📘 DataStax Bulk Loader with Astra

Use DataStax Bulk Loader (dsbulk) to load and unload data in CSV or JSON format with your DataStax Astra DB database efficiently and reliably.

You can use dsbulk as a standalone tool to remotely connect to a cluster. The tool is not required to run locally on an instances, but can be used in this configuration.

B - Prerequisites

This article was written for DataStax Bulk Loader version 1.9.1.

Starting with version 1.9, dsbulk can detect and respect server-side rate limiting. This is very useful when working with Astra DB, which by default has some throughput guardrails in place.

C - Installation

✅ Step 1 : Download the archive and unzip locally

curl -OL \
          && tar xvzf dsbulk-1.9.1.tar.gz \
          && rm -f dsbulk-1.9.1.tar.gz

it will take a few seconds (file is about 30M)...

  % Total    % Received % Xferd  Average Speed   Time    Time     Time  Current
                                 Dload  Upload   Total   Spent    Left  Speed
 49 30.0M   49 14.8M    0     0   343k      0  0:01:29  0:00:44  0:00:45  244k

D - Usage

📘 Load Data

  • Given a table
CREATE TABLE better_reads.book_by_id (
    id text PRIMARY KEY,
    author_id list<text>,
    author_names list<text>,
    book_description text,
    book_name text,
    cover_ids list<text>,
    published_date date
  • A sample CSV could be:
1234|["id1","id2","id3"]|["name1","name2","name3"]|this is a dsecription|Book name|["cover1","cover2"]|2022-02-02
  • Loaded with the following command:
dsbulk load \
    -url book_by_id.csv \
    -c csv \
    -delim '|' \
    -k better_reads \
    -t book_by_id \
    --schema.allowMissingFields true \
    -u clientId \
    -p clientSecret \

📘 Export Data

  • Unloaded the same table with the following command:
dsbulk unload \
    -k better_reads \
    -t book_by_id \
    -c csv \
    -u clientId \
    -p clientSecret \
    -b \
    > book_by_id_export.csv

📘 Count Table Records

  • Counted the rows in the table with the following command:
dsbulk count \
    -k better_reads \
    -t book_by_id \
    -u clientId \
    -p clientSecret \
  • Produces the following output:
Operation directory: /local/dsbulk-1.9.1/logs/COUNT_20220223-213637-046128
  total | failed | rows/s |  p50ms |  p99ms | p999ms
143,475 |      0 | 87,509 | 155.34 | 511.71 | 511.71
Operation COUNT_20220223-213637-046128 completed successfully in 1 second.

Last update: 2022-12-15