Apache Cassandra features supported by Azure Cosmos DB for Apache Cassandra
APPLIES TO: Cassandra
Azure Cosmos DB is Microsoft's globally distributed multi-model database service. You can communicate with the Azure Cosmos DB for Apache Cassandra through the Cassandra Query Language (CQL) Binary Protocol v4 wire protocol compliant open-source Cassandra client drivers.
By using the Azure Cosmos DB for Apache Cassandra, you can enjoy the benefits of the Apache Cassandra APIs and the enterprise capabilities that Azure Cosmos DB provides. The enterprise capabilities include global distribution, automatic scale out partitioning, availability and latency guarantees, encryption at rest, backups, and much more.
Cassandra protocol
The Azure Cosmos DB for Apache Cassandra is compatible with Cassandra Query Language (CQL) v3.11 API (backward-compatible with version 2.x). The supported CQL commands, tools, limitations, and exceptions are listed below. Any client driver that understands these protocols should be able to connect to Azure Cosmos DB for Apache Cassandra.
Azure Managed Instance for Apache Cassandra
For some customers, adapting to API for Cassandra can be a challenge due to differences in behaviour and/or configuration, especially for lift-and-shift migrations. If a feature that is critical for your application is listed as not supported below, consider using Azure Managed Instance for Apache Cassandra. This is a first-party Azure service for hosting and maintaining pure open-source Apache Cassandra clusters with 100% compatibility.
Cassandra driver
The following versions of Cassandra drivers are supported by Azure Cosmos DB for Apache Cassandra:
CQL data types
Azure Cosmos DB for Apache Cassandra supports the following CQL data types:
Type | Supported |
---|---|
ascii |
Yes |
bigint |
Yes |
blob |
Yes |
boolean |
Yes |
counter |
Yes |
date |
Yes |
decimal |
Yes |
double |
Yes |
float |
Yes |
frozen |
Yes |
inet |
Yes |
int |
Yes |
list |
Yes |
set |
Yes |
smallint |
Yes |
text |
Yes |
time |
Yes |
timestamp |
Yes |
timeuuid |
Yes |
tinyint |
Yes |
tuple |
Yes |
uuid |
Yes |
varchar |
Yes |
varint |
Yes |
tuples |
Yes |
udts |
Yes |
map |
Yes |
Static is supported for data type declaration.
CQL functions
Azure Cosmos DB for Apache Cassandra supports the following CQL functions:
Command | Supported |
---|---|
Token * |
Yes |
ttl *** |
Yes |
writetime *** |
Yes |
cast ** |
Yes |
Note
* API for Cassandra supports token as a projection/selector, and only allows token(pk) on the left-hand side of a where clause. For example, WHERE token(pk) > 1024
is supported, but WHERE token(pk) > token(100)
is not supported.
** The cast()
function is not nestable in API for Cassandra. For example, SELECT cast(count as double) FROM myTable
is supported, but SELECT avg(cast(count as double)) FROM myTable
is not supported.
*** Custom timestamps and TTL specified with the USING
option are applied at a row level (and not per cell).
Aggregate functions:
Command | Supported |
---|---|
avg |
Yes |
count |
Yes |
min |
Yes |
max |
Yes |
sum |
Yes |
Note
Aggregate functions work on regular columns, but aggregates on clustering columns are not supported.
Blob conversion functions:
Command | Supported |
---|---|
typeAsBlob(value) |
Yes |
blobAsType(value) |
Yes |
UUID and timeuuid functions:
Command | Supported |
---|---|
dateOf() |
Yes |
now() |
Yes |
minTimeuuid() |
Yes |
unixTimestampOf() |
Yes |
toDate(timeuuid) |
Yes |
toTimestamp(timeuuid) |
Yes |
toUnixTimestamp(timeuuid) |
Yes |
toDate(timestamp) |
Yes |
toUnixTimestamp(timestamp) |
Yes |
toTimestamp(date) |
Yes |
toUnixTimestamp(date) |
Yes |
CQL commands
Azure Cosmos DB supports the following database commands on API for Cassandra accounts.
Command | Supported |
---|---|
ALLOW FILTERING |
Yes |
ALTER KEYSPACE |
N/A (PaaS service, replication managed internally) |
ALTER MATERIALIZED VIEW |
Yes |
ALTER ROLE |
No |
ALTER TABLE |
Yes |
ALTER TYPE |
No |
ALTER USER |
No |
BATCH |
Yes (unlogged batch only) |
COMPACT STORAGE |
N/A (PaaS service) |
CREATE AGGREGATE |
No |
CREATE CUSTOM INDEX (SASI) |
No |
CREATE INDEX |
Yes (including named indexes but full FROZEN collection is not supported) |
CREATE FUNCTION |
No |
CREATE KEYSPACE (replication settings ignored) |
Yes |
CREATE MATERIALIZED VIEW |
Yes |
CREATE TABLE |
Yes |
CREATE TRIGGER |
No |
CREATE TYPE |
Yes |
CREATE ROLE |
No |
CREATE USER (Deprecated in native Apache Cassandra) |
No |
DELETE |
Yes |
DISTINCT |
No |
DROP AGGREGATE |
No |
DROP FUNCTION |
No |
DROP INDEX |
Yes |
DROP KEYSPACE |
Yes |
DROP MATERIALIZED VIEW |
Yes |
DROP ROLE |
No |
DROP TABLE |
Yes |
DROP TRIGGER |
No |
DROP TYPE |
Yes |
DROP USER (Deprecated in native Apache Cassandra) |
No |
GRANT |
No |
INSERT |
Yes |
LIST PERMISSIONS |
No |
LIST ROLES |
No |
LIST USERS (Deprecated in native Apache Cassandra) |
No |
REVOKE |
No |
SELECT |
Yes |
UPDATE |
Yes |
TRUNCATE |
Yes |
USE |
Yes |
Lightweight Transactions (LWT)
Component | Supported |
---|---|
DELETE IF EXISTS |
Yes |
DELETE conditions |
Yes |
INSERT IF NOT EXISTS |
Yes |
UPDATE IF EXISTS |
Yes |
UPDATE IF NOT EXISTS |
Yes |
UPDATE conditions |
Yes |
Note
Lightweight transactions currently aren't supported for accounts that have multi-region writes enabled.
CQL Shell commands
Azure Cosmos DB supports the following database commands on API for Cassandra accounts.
Command | Supported |
---|---|
CAPTURE |
Yes |
CLEAR |
Yes |
CONSISTENCY * |
N/A |
COPY |
No |
DESCRIBE |
Yes |
cqlshExpand |
No |
EXIT |
Yes |
LOGIN |
N/A (CQL function USER is not supported, hence LOGIN is redundant) |
PAGING |
Yes |
SERIAL CONSISTENCY * |
N/A |
SHOW |
Yes |
SOURCE |
Yes |
TRACING |
N/A (API for Cassandra is backed by Azure Cosmos DB - use diagnostic logging for troubleshooting) |
Note
Consistency works differently in Azure Cosmos DB, see here for more information.
JSON Support
Command | Supported |
---|---|
SELECT JSON |
Yes |
INSERT JSON |
Yes |
fromJson() |
No |
toJson() |
No |
API for Cassandra limits
Azure Cosmos DB for Apache Cassandra does not have any limits on the size of data stored in a table. Hundreds of terabytes or Petabytes of data can be stored while ensuring partition key limits are honored. Similarly, every entity or row equivalent does not have any limits on the number of columns. However, the total size of the entity should not exceed 2 MB. The data per partition key cannot exceed 20 GB as in all other APIs.
Tools
Azure Cosmos DB for Apache Cassandra is a managed service platform. The platform does not require any management overhead or utilities such as Garbage Collector, Java Virtual Machine(JVM), and nodetool to manage the cluster. Tools such as cqlsh that utilizes Binary CQLv4 compatibility are supported.
- Azure portal's data explorer, metrics, log diagnostics, PowerShell, and CLI are other supported mechanisms to manage the account.
CQL shell
You can connect to the API for Cassandra in Azure Cosmos DB by using the CQLSH installed on a local machine. It comes with Apache Cassandra 3.11 and works out of the box by setting the environment variables. The following sections include the instructions to install, configure, and connect to API for Cassandra in Azure Cosmos DB, on Windows or Linux using CQLSH.
Warning
Connections to Azure Cosmos DB for Apache Cassandra will not work with DataStax Enterprise (DSE) or Cassandra 4.0 versions of CQLSH. Please ensure you use only v3.11 open source Apache Cassandra versions of CQLSH when connecting to API for Cassandra.
Windows:
- Install Python 3
- Install PIP
- Before install PIP, download the get-pip.py file.
- Launch a command prompt if it isn't already open. To do so, open the Windows search bar, type cmd and select the icon.
- Then, run the following command to download the get-pip.py file:
curl https://bootstrap.pypa.io/get-pip.py -o get-pip.py
- Install PIP on Windows
python get-pip.py
- Verify the PIP installation (look for a message from step 3 to confirm which folder PIP was installed in and then navigate to that folder and run the command pip help).
- Install CQLSH using PIP
pip3 install cqlsh==5.0.3
- Install Python 2
- Run the CQLSH using the authentication mechanism.
Note
You would need to set the environment variables to point to the Python 2 folder.
Install on Unix/Linux/Mac:
# Install default-jre and default-jdk
sudo apt install default-jre
sudo apt-get update
sudo apt install default-jdk
# Import the Baltimore CyberTrust root certificate:
curl https://cacert.omniroot.com/bc2025.crt > bc2025.crt
keytool -importcert -alias bc2025ca -file bc2025.crt
# Install the Cassandra libraries in order to get CQLSH:
echo "deb https://downloads.apache.org/cassandra/debian 311x main" | sudo tee -a /etc/apt/sources.list.d/cassandra.sources.list
curl https://downloads.apache.org/cassandra/KEYS | sudo apt-key add -
sudo apt-get update
sudo apt-get install cassandra=3.11.13
Connect with Unix/Linux/Mac:
# Export the SSL variables:
export SSL_VERSION=TLSv1_2
export SSL_VALIDATE=false
# Connect to Azure Cosmos DB for Apache Cassandra:
cqlsh <YOUR_ACCOUNT_NAME>.cassandra.cosmosdb.azure.com 10350 -u <YOUR_ACCOUNT_NAME> -p <YOUR_ACCOUNT_PASSWORD> --ssl --protocol-version=4
Connect with Docker:
docker run -it --rm -e SSL_VALIDATE=false -e SSL_VERSION=TLSv1_2 cassandra:3.11 cqlsh <account_name>.cassandra.cosmos.azure.com 10350 -u <YOUR_ACCOUNT_NAME> -p <YOUR_ACCOUNT_PASSWORD> --ssl
All CRUD operations that are executed through a CQL v4 compatible SDK will return extra information about error and request units consumed. The DELETE and UPDATE commands should be handled with resource governance taken into consideration, to ensure the most efficient use of the provisioned throughput.
- Note gc_grace_seconds value must be zero if specified.
var tableInsertStatement = table.Insert(sampleEntity);
var insertResult = await tableInsertStatement.ExecuteAsync();
foreach (string key in insertResult.Info.IncomingPayload)
{
byte[] valueInBytes = customPayload[key];
double value = Encoding.UTF8.GetString(valueInBytes);
Console.WriteLine($"CustomPayload: {key}: {value}");
}
Consistency mapping
Azure Cosmos DB for Apache Cassandra provides choice of consistency for read operations. The consistency mapping is detailed here.
Permission and role management
Azure Cosmos DB supports Azure role-based access control (Azure RBAC) for provisioning, rotating keys, viewing metrics and read-write and read-only passwords/keys that can be obtained through the Azure portal. Azure Cosmos DB does not support roles for CRUD activities.
Keyspace and Table options
The options for region name, class, replication_factor, and datacenter in the "Create Keyspace" command are ignored currently. The system uses the underlying Azure Cosmos DB's global distribution replication method to add the regions. If you need the cross-region presence of data, you can enable it at the account level with PowerShell, CLI, or portal, to learn more, see the how to add regions article. Durable_writes can't be disabled because Azure Cosmos DB ensures every write is durable. In every region, Azure Cosmos DB replicates the data across the replica set that is made up of four replicas and this replica set configuration can't be modified.
All the options are ignored when creating the table, except gc_grace_seconds, which should be set to zero. The Keyspace and table have an extra option named "cosmosdb_provisioned_throughput" with a minimum value of 400 RU/s. The Keyspace throughput allows sharing throughput across multiple tables and it is useful for scenarios when all tables are not utilizing the provisioned throughput. Alter Table command allows changing the provisioned throughput across the regions.
CREATE KEYSPACE sampleks WITH REPLICATION = { 'class' : 'SimpleStrategy'} AND cosmosdb_provisioned_throughput=2000;
CREATE TABLE sampleks.t1(user_id int PRIMARY KEY, lastname text) WITH cosmosdb_provisioned_throughput=2000;
ALTER TABLE gks1.t1 WITH cosmosdb_provisioned_throughput=10000 ;
Secondary Index
API for Cassandra supports secondary indexes on all data types except frozen collection types, decimal, and variant types.
Usage of Cassandra retry connection policy
Azure Cosmos DB is a resource governed system. You can do a certain number of operations in a given second based on the request units consumed by the operations. If an application exceeds that limit in a given second, requests are rate-limited and exceptions will be thrown. The API for Cassandra in Azure Cosmos DB translates these exceptions to overloaded errors on the Cassandra native protocol. To ensure that your application can intercept and retry requests in case of rate limitation, the spark and the Java extensions are provided. See also Java code samples for version 3 and version 4 Datastax drivers, when connecting to API for Cassandra in Azure Cosmos DB. If you use other SDKs to access API for Cassandra in Azure Cosmos DB, create a retry policy to retry on these exceptions. Alternatively, enable server-side retries for API for Cassandra.
Next steps
- Get started with creating a API for Cassandra account, database, and a table by using a Java application