Cassandra Query Language (CQL) v2.0

Table of Contents

  1. Cassandra Query Language (CQL) v2.0
    1. Table of Contents
    2. Syntax conventions
    3. Overall syntax
    4. Data Storage Types
      1. Working with dates
    5. USE
    6. SELECT
      1. Specifying Columns
      2. Column Family
      3. Consistency Level
      4. Filtering rows
      5. Limits
    7. INSERT
    8. UPDATE
      1. Column Family
      2. Consistency Level
      3. Timestamp
      4. TTL
      5. Specifying Columns and Row
      6. Updating Counter Columns
    9. DELETE
      1. Specifying Columns
      2. Column Family
      3. Consistency Level
      4. Specifying Rows
    10. TRUNCATE
    11. BATCH
      1. Specifying Key Type
      2. Specifying Column Types (optional)
      3. Column Family Options (optional)
    17. DROP INDEX
      1. Changing the type of a typed column
      2. Adding a typed column
      3. Dropping a typed column
    19. Common Idioms
      1. Specifying Consistency
  2. Versioning
  3. Changes

Syntax conventions

To aid in specifying the CQL syntax, we will use the following conventions in this document:

<Start> ::= <CQL_Statement>*
SELECT sample_usage FROM cql;

Overall syntax

CQL consists of statements. As in SQL, some statements directly make changes to data, some look up data, and some change the way data is stored.

All statements end with a semicolon.

<CQL_Statement> ::= <statementBody> ";"
<statementBody> ::= <useStatement>
                  | <selectStatement>
                  | <dataChangeStatement>
                  | <schemaChangeStatement>
<dataChangeStatement> ::= <insertStatement>
                        | <updateStatement>
                        | <batchStatement>
                        | <deleteStatement>
                        | <truncateStatement>
<schemaChangeStatement> ::= <createKeyspaceStatement>
                          | <createColumnFamilyStatement>
                          | <createIndexStatement>
                          | <dropKeyspaceStatement>
                          | <dropColumnFamilyStatement>
                          | <dropIndexStatement>
                          | <alterTableStatement>

String literals and identifiers (including keyspace and column family names) are case-sensitive, but CQL keywords are not. We show CQL keywords in this document in UPPERCASE merely as a convention to aid readability.

Literal values can be expressed in several ways in CQL.

<term> ::= "KEY"
         | <identifier>
         | <stringLiteral>
         | <integer>
         | <float>
         | <uuid>

Comments in CQL can begin with a double dash (--) or a double slash (//) and extend to the end of the line. Multiline comments are enclosed in /* ... */.

Data Storage Types


<storageType> ::= "ascii"
                | "bigint"
                | "blob"
                | "boolean"
                | "counter"
                | "decimal"
                | "double"
                | "float"
                | "int"
                | "text"
                | "timestamp"
                | "uuid"
                | "varchar"
                | "varint"

The following table gives additional information on the available data types.

asciiASCII character string
bigint64-bit signed long
blobArbitrary bytes (no validation)
booleantrue or false
counterCounter column (64-bit long)
decimalVariable-precision decimal
double64-bit IEEE-754 floating point
float32-bit IEEE-754 floating point
int32-bit signed int
textUTF8 encoded string
timestampA timestamp. See Working with dates below for more information.
uuidType 1 or type 4 UUID
varcharUTF8 encoded string
varintArbitrary-precision integer

Note: In addition to the recognized types listed above, it is also possible to supply a string containing the name of a class (a sub-class of AbstractType loadable by Cassandra). The class name should either be fully qualified, or relative to the org.apache.cassandra.db.marshal package.

Working with dates

Values serialized with the timestamp type are encoded as 64-bit signed integers representing a number of milliseconds since the standard base time known as “the epoch”: January 1 1970 at 00:00:00 GMT.

Timestamp types can be input in CQL as simple long integers, giving the number of milliseconds since the epoch, as defined above.

Timestamp types can also be input as string literals in any of the following ISO 8601 formats, each representing the time and date Jan 2, 2003, at 04:05:00 AM, GMT.:

The +0000 above is an RFC 822 4-digit time zone specification; +0000 refers to GMT. US Pacific Standard Time is -0800. The time zone may be omitted if desired— the date will be interpreted as being in the time zone under which the coordinating Cassandra node is configured.

There are clear difficulties inherent in relying on the time zone configuration being as expected, though, so it is recommended that the time zone always be specified for timestamps when feasible.

The time of day may also be omitted, if the date is the only piece that matters:

In that case, the time of day will default to 00:00:00, in the specified or default time zone.



<useStatement> ::= "USE" <term>


USE myApp;

A USE statement consists of the USE keyword, followed by a valid keyspace name. Its purpose is to assign the per-connection, current working keyspace. All subsequent keyspace-specific actions will be performed in the context of the keyspace selected, unless otherwise specified, until another USE statement is issued or the connection terminates.



<selectStatement> ::= "SELECT" <whatToSelect>
                        "FROM" ( <name> "." )? <name>
                               ( "USING" "CONSISTENCY" <consistencylevel> )?
                               ( "WHERE" <selectWhereClause> )?
                               ( "LIMIT" <integer> )?
<whatToSelect> ::= <term> ( "," <term> )*
                 | ("FIRST" <integer> )? "REVERSED"? <columnRange>
                 | "COUNT" "(" <countTarget> ")"
<columnRange> ::= <term> ".." <term>
                | "*"
<countTarget> ::= "*"
                | "1"
<name> ::= <identifier>
         | <stringLiteral>
         | <integer>
<selectWhereClause> ::= <relation> ( "AND" <relation> )*
                      | <term> "IN" "(" <term> ( "," <term> )* ")"
<relation> ::= <term> <relationOperator> <term>
<relationOperator> ::= "=" | "<" | ">" | "<=" | ">="


SELECT Name, Occupation FROM People WHERE key IN (199, 200, 207);
SELECT FIRST 3 REVERSED 'time199'..'time100' FROM Events;
SELECT COUNT(*) FROM system.Migrations;

A SELECT is used to read one or more records from a Cassandra column family. It returns a result-set of rows, where each row consists of a key and a collection of columns corresponding to the query.

Specifying Columns

SELECT col1, col2 FROM ...
SELECT range_lo..range_hi FROM ...
SELECT FIRST 4 REVERSED range_hi..range_lo FROM ...

The SELECT expression determines which columns will appear in the results and can take a few different forms, as shown above. The simplest is a comma-separated list of column names. Note that column names in Cassandra can be specified with string literals or integers, in addition to identifiers.

It is also possible to specify a range of column names. The range notation consists of start and end column names, separated by two periods (..). The set of columns returned for a range is start and end inclusive. A single star (*) may be used as a range to request “all columns”.

When using a range, it is sometimes useful to limit the number of columns that can be returned as part of each row (since Cassandra is schemaless, it is not necessarily possible to determine ahead of time how many columns will be in the result set). To accomplish this, use the FIRST clause with an integer to specify an upper limit on the number of columns returned per row. The default limit is 10,000 columns.

The REVERSED option causes the sort order of the columns returned to be reversed. This affects the FIRST clause; when limiting the columns returned, the columns at the end of the range will be selected instead of the ones at the beginning of the range.

A SELECT expression may also be COUNT(*). In this case, the result will be only one value: the number of rows which matched the query.

It is worth noting that unlike the projection in a SQL SELECT, there is no guarantee that the results will contain all of the columns specified, because Cassandra is schemaless.

Column Family

SELECT ... FROM MyApp.LocationSnapshots ...;
SELECT ... FROM EventTimeline ...;

The FROM clause is used to specify the Cassandra column family applicable to a SELECT query. Unlike other operations on column families, the keyspace in which the column family exists may also be specified by giving its name before the column family name, and separating them by a dot (.). If the keyspace is not specified, the current keyspace will be used, as per normal.

Consistency Level


Following the column family clause is an optional consistency level specification.

Filtering rows

SELECT ... WHERE KEY = 11194251 AND startdate = '2011-10-08-0500';
SELECT ... WHERE KEY >= 'AM' and KEY =< 'AZ' AND module = 17;
SELECT ... WHERE keyalias IN ('key1', 'key2', 'key3', ...);

The WHERE clause provides for filtering the rows that appear in results. The clause can filter on a key name, or range of keys, and in the case of indexed columns, on column values. Key filters are specified using the KEY keyword or key alias name, followed by a relational operator (one of =, >, >=, <, and <=), and then a term value. When terms appear on both sides of a relational operator it is assumed the filter applies to an indexed column. With column index filters, the term on the left of the operator must be the name of the indexed column, and the term on the right is the value to filter on.

Note: The greater-than and less-than operators (> and <) result in key ranges that are inclusive of the terms. There is no supported notion of “strictly” greater-than or less-than; these operators are merely supported as aliases to >= and <=.


SELECT ... WHERE favoriteArtist = 'The Mighty Mighty Bosstones' LIMIT 90000;

The LIMIT option to a SELECT expression limits the number of rows returned by a query. LIMIT defaults to 10,000 when left unset.



<insertStatement> ::= "INSERT" "INTO" <name>
                               "(" <term> "," <term> ( "," <term> )* ")"
                      "VALUES" "(" <term> "," <term> ( "," <term> )* ")"
                      ( "USING" <usingOption> ( "AND" <usingOption> )* )?
<usingOption> ::= "CONSISTENCY" <consistencylevel>
                | "TIMESTAMP" <integer>
                | "TTL" <integer>


INSERT INTO NerdMovies (KEY, 11924)
                VALUES ('Serenity', 'Nathan Fillion')

An INSERT is used to write one or more columns to a record in a Cassandra column family. No results are returned.

The first column name in the INSERT list must be the name of the column family key. Also, there must be more than one column name specified (Cassandra rows are not considered to exist with only a key and no associated columns).

Unlike in SQL, the semantics of INSERT and UPDATE are identical. In either case a record is created if none existed before, and updated when it does. For information on query modifiers and types, see the UPDATE section below.



<updateStatement> ::= "UPDATE" <name>
                        ( "USING" <usingOption> ( "AND" <usingOption> )* )?
                        "SET" <assignment> ( "," <assignment> )*
                        "WHERE" <updateWhereClause>
<assignment> ::= <term> "=" <term>
               | <term> "=" <term> "+" <term>
               | <term> "=" <term> "-" <term>
<updateWhereClause> ::= <term> "=" <term>
                      | <term> "IN" "(" <term> ( "," <term> )* ")"


       SET 'A 1194' = 'The Empire Strikes Back',
           'B 1194' = 'Han Solo'
     WHERE KEY = B70DE1D0-9908-4AE3-BE34-5573E5B09F14;
UPDATE UserActionCounts SET total = total + 2 WHERE keyalias = 523;

An UPDATE is used to write one or more columns to a record in a Cassandra column family. No results are returned. The row key can be specified using the KEY keyword or by a key alias set per column family.

Column Family

Statements begin with the UPDATE keyword followed by a Cassandra column family name.

Consistency Level


Following the column family identifier is an optional USING clause, which can specify the consistency level for the update, or the timestamp and/or the TTL for the new columns.


UPDATE Foo USING TIMESTAMP=1318452291034 ...

UPDATE supports setting client-supplied optional timestamp for modification.



UPDATE supports setting a time to live (TTL), in seconds, for each of the added columns.

Specifying Columns and Row

UPDATE ... SET col1 = val1, col2 = val2 WHERE KEY = key1;
UPDATE ... SET col3 = val3 WHERE KEY IN (key1, key2, key3);
UPDATE ... SET col4 = 22 WHERE keyalias = key4;

Rows are created or updated by supplying column names and values, after the SET keyword, in term assignment format. Multiple columns can be set by separating the name/value pairs using commas. Each update statement requires a precise set of row keys to be specified using a WHERE clause and the KEY keyword or key alias.

Updating Counter Columns

UPDATE ... SET name1 = name1 + <value> ...;
UPDATE ... SET name1 = name1 - <value> ...;

Counter columns can be incremented or decremented by an arbitrary numeric value though the assignment of an expression that adds or substracts the value.



<deleteStatement> ::= "DELETE" ( <term> ( "," <term> )* )?
                        "FROM" <name>
                        ( "USING" <deleteOption> ( "AND" <deleteOption> )* )?
                        "WHERE" <updateWhereClause>
<deleteOption> ::= "CONSISTENCY" <consistencylevel>
                 | "TIMESTAMP" <integer>


DELETE col1, col2, col3 FROM Planeteers USING CONSISTENCY ONE WHERE KEY = 'Captain';
DELETE FROM MastersOfTheUniverse WHERE KEY IN ('Man-At-Arms', 'Teela');

A DELETE is used to perform the removal of one or more columns from one or more rows. The key can be given using the KEY keyword or by the key alias set per column family.

Specifying Columns

Following the DELETE keyword is an optional comma-delimited list of column name terms. When no column names are specified, the remove applies to the entire row(s) matched by the WHERE clause.

Column Family

The column family name follows the list of column names and the keyword FROM.

Consistency Level

Following the column family identifier is an optional consistency level specification.

Specifying Rows

DELETE ... WHERE KEY = 'some_key_value';
DELETE ... WHERE keyalias IN (key1, key2);

The WHERE clause is used to determine to which row(s) a DELETE applies. The first form allows the specification of a single keyname using the KEY keyword (or by key alias) and the = operator. The second form allows a list of keyname terms to be specified using the IN notation and a parenthesized list of comma-delimited keyname terms.



<truncateStatement> ::= "TRUNCATE" <name>


TRUNCATE super_important_data;

TRUNCATE accepts a single argument for the column family name, and permanently removes all data from said column family.



<batchStatement> ::= "BEGIN" "BATCH"
                        ( "USING" <usingOption> ( "AND" <usingOption> )* )?
                        <batchStatementMember> ( ";" <batchStatementMember> )*
                     "APPLY" "BATCH"
<batchStatementMember> ::= <insertStatement>
                         | <updateStatement>
                         | <deleteStatement>


  INSERT INTO users (KEY, password, name) VALUES ('user2', 'ch@ngem3b', 'second user')
  UPDATE users SET password = 'ps22dhds' WHERE KEY = 'user2'
  INSERT INTO users (KEY, password) VALUES ('user3', 'ch@ngem3c')
  DELETE name FROM users WHERE key = 'user2'
  INSERT INTO users (KEY, password, name) VALUES ('user4', 'ch@ngem3c', 'Andrew')

BATCH supports setting a client-supplied optional global timestamp which will be used for each of the operations included in the batch.

A single consistency level is used for the entire batch. It appears after the BEGIN BATCH statement, and uses the standard consistency level specification. Batched statements default to CONSISTENCY.ONE when left unspecified.

Only data modification statements (specifically, UPDATE, INSERT, and DELETE) are allowed in a BATCH statement. BATCH is not an analogue for SQL transactions.

NOTE: While there are no isolation guarantees, UPDATE queries are atomic within a given record.



<createKeyspaceStatement> ::= "CREATE" "KEYSPACE" <name>
                                 "WITH" <optionName> "=" <optionVal>
                                 ( "AND" <optionName> "=" <optionVal> )*
<optionName> ::= <identifier>
               | <optionName> ":" <identifier>
               | <optionName> ":" <integer>
<optionVal> ::= <stringLiteral>
              | <identifier>
              | <integer>


CREATE KEYSPACE Excelsior WITH strategy_class = 'SimpleStrategy'
    AND strategy_options:replication_factor = 1;
CREATE KEYSPACE Excalibur WITH strategy_class = 'NetworkTopologyStrategy'
    AND strategy_options:DC1 = 1 AND strategy_options:DC2 = 3;

The CREATE KEYSPACE statement creates a new top-level namespace (aka “keyspace”). Valid names are any string constructed of alphanumeric characters and underscores. Names which do not work as valid identifiers or integers should be quoted as string literals. Properties such as replication strategy and count are specified during creation using the following accepted keyword arguments:

strategy_classyesThe name of the replication strategy class which should be used for the new keyspace. Some often-used classes are SimpleStrategy and NetworkTopologyStrategy.
strategy_optionsnoMost strategies require additional arguments which can be supplied by appending the option name to the strategy_options keyword, separated by a colon (:). For example, a strategy option of “DC1” with a value of “1” would be specified as strategy_options:DC1 = 1; replication_factor for SimpleStrategy could be strategy_options:replication_factor=3.



<createColumnFamilyStatement> ::= "CREATE" "COLUMNFAMILY" <name>
                                    "(" <term> <storageType> "PRIMARY" "KEY"
                                        ( "," <term> <storageType> )* ")"
                                   ( "WITH" <identifier> "=" <cfOptionVal>
                                     ( "AND" <identifier> "=" <cfOptionVal> )* )?
<cfOptionVal> ::= <storageType>
                | <identifier>
                | <stringLiteral>
                | <integer>
                | <float>


    WITH comparator=timestamp AND default_validation=int;
    species text,
    alias text,
    population varint
) WITH comment='Important biological records'
   AND read_repair_chance = 1.0;

CREATE COLUMNFAMILY statements create new column family namespaces under the current keyspace. Valid column family names are strings of alphanumeric characters and underscores, which begin with a letter.

Specifying Key Type

CREATE ... (KEY ascii PRIMARY KEY, ... ) ...

When creating a new column family, you must specify the key type. The list of possible types is identical to column comparators/validators (see Data Storage Types), except it probably does not make sense to use counter for a key. It’s important to note that the key type you use must be compatible with the partitioner in use. For example, OrderPreservingPartitioner and CollatingOrderPreservingPartitioner both require UTF-8 keys. If you use an identifier for the primary key name, instead of the KEY keyword, a key alias will be set automatically.

Specifying Column Types (optional)

CREATE ... ( ... , name1 type1, name2 type2, ... ) ...

It is possible to assign columns a type during column family creation. Columns configured with a type are validated accordingly when a write occurs, and intelligent CQL drivers and interfaces will be able to decode the column values correctly when receiving them. Column types are specified as a parenthesized, comma-separated list of column term and type pairs. See Data Storage Types for the list of recognized types.

Column Family Options (optional)

CREATE COLUMNFAMILY ... WITH keyword1 = arg1 AND keyword2 = arg2;

A number of optional keyword arguments can be supplied to control the configuration of a new column family.

comparatortextDetermines the storage type of column names (which itself determines the sorting and validation of column names). Valid values are listed in the Data Storage Types table above.
commentnoneA free-form, human-readable comment.
row_cache_providerSerializingCacheProvider if JNA is present, otherwise ConcurrentHashMapCacheProviderA factory for the cache with which to back the row cache.
row_cache_size0Number of rows whose entire contents to cache in memory.
key_cache_size200000Number of keys per SSTable whose locations are kept in memory in “mostly LRU” order.
read_repair_chance1.0The probability with which read repairs should be invoked on non-quorum reads.
gc_grace_seconds864000Time to wait before garbage collecting tombstones (deletion markers).
default_validationtextDetermines the default storage type of column values (which itself determines the validation for column values). This option does not affect the types of columns which were defined in a CREATE COLUMNFAMILY statement— only new columns. Valid values are listed in the Data Storage Types table above.
min_compaction_threshold4Minimum number of SSTables needed to start a minor compaction.
max_compaction_threshold32Maximum number of SSTables allowed before a minor compaction is forced.
row_cache_save_period_in_seconds0Number of seconds between saving row caches.
key_cache_save_period_in_seconds14400Number of seconds between saving key caches.



<createIndexStatement> ::= "CREATE" "INDEX" <identifier>? "ON"
                               <name> "(" <term> ")"


CREATE INDEX userIndex ON NerdMovies (user);
CREATE INDEX ON Mutants (abilityId);

A CREATE INDEX statement is used to create a new, automatic secondary index on the given column family, for the named column. A name for the index itself can be specified before the ON keyword, if desired. A single column name must be specified inside the parentheses. It is not necessary for the column to exist on any current rows (Cassandra is schemaless), but the column must already have a type (specified during the CREATE COLUMNFAMILY, or added afterwards with ALTER COLUMNFAMILY.



<dropKeyspaceStatement> ::= "DROP" "KEYSPACE" <name>



A DROP KEYSPACE statement results in the immediate, irreversible removal of a keyspace, including all column families in it, and all data contained in those column families.



<dropColumnFamilyStatement> ::= "DROP" "COLUMNFAMILY" <name>


DROP COLUMNFAMILY worldSeriesAttendees;

A DROP COLUMNFAMILY statement results in the immediate, irreversible removal of a column family, including all data contained in it.



<dropIndexStatement> ::= "DROP" "INDEX" <name>


DROP INDEX cf_col_idx;

A DROP INDEX statement is used to drop an existing secondary index.



<alterTableStatement> ::= "ALTER" "COLUMNFAMILY" <name> <alterInstructions>
<alterInstructions> ::= "ALTER" <name> "TYPE" <storageType>
                      | "ADD" <name> <storageType>
                      | "DROP" <name>

An ALTER statement is used to manipulate column family column metadata. It allows you to add new columns, drop existing columns, or change the data storage type of existing columns. No results are returned.

Specify the name of the column family to be changed after the ALTER COLUMNFAMILY keywords, and the name of the column to be changed, added, or dropped after the keyword corresponding to the type of change desired (ALTER, ADD, DROP).

Changing the type of a typed column

ALTER COLUMNFAMILY addamsFamily ALTER lastKnownLocation TYPE uuid;

ALTER COLUMNFAMILY ... ALTER changes the expected storage type for a column. The column must already have a type in the column family metadata. The column may or may not already exist in current rows— but be aware that no validation of existing data is done. The bytes stored in values for that column will remain unchanged, and if existing data is not deserializable according to the new type, this may cause your CQL driver or interface to report errors.

Adding a typed column

ALTER COLUMNFAMILY addamsFamily ADD gravesite varchar;

The ALTER COLUMNFAMILY ... ADD variant adds a typed column to a column family. The column must not already have a type in the column family metadata. The same warnings from the above ALTER section, about there being no validation of existing data, apply here as well.

Dropping a typed column

ALTER COLUMNFAMILY addamsFamily DROP gender;

An ALTER COLUMNFAMILY ... DROP statement removes the type of a column from the column family metadata. Note that this does not remove the column from current rows; it just removes the metadata saying that the bytes stored under that column are expected to be deserializable according to a certain type.

Common Idioms

Specifying Consistency

<consistency> ::= "ANY"
                | "ONE"
                | "QUORUM"
                | "ALL"
                | "LOCAL_QUORUM"
                | "EACH_QUORUM"

Consistency level specifications are made up the keywords USING CONSISTENCY, followed by a consistency level identifier. Valid consistency level identifiers are as listed above. When not specified, USING CONSISTENCY ONE is the default.

Consult your Cassandra documentation for information about how consistency levels work.


Versioning of the CQL language adheres to the Semantic Versioning guidelines. Versions take the form X.Y.Z where X, Y, and Z are integer values representing major, minor, and patch level respectively. There is no correlation between Cassandra release versions and the CQL language version.

MajorThe major version must be bumped when backward incompatible changes are introduced. This should rarely (if ever) occur.
MinorMinor version increments occur when new, but backward compatible, functionality is introduced.
PatchThe patch version is incremented when bugs are fixed.


Wed, 12 Oct 2011 16:53:00 -0500 - paul cannon
 * Rework whole doc, adding syntax specifics and additional explanations

Fri, 09 Sep 2011 11:43:00 -0500 - Jonathan Ellis
 * add int data type

Wed, 07 Sep 2011 09:01:00 -0500 - Jonathan Ellis
 * Updated version to 2.0; Documented row-based count()
 * Updated list of supported data types
Wed, 10 Aug 2011 11:22:00 -0500 - Eric Evans
 * Improved INSERT vs. UPDATE wording.
 * Documented counter column incr/descr.
Sat, 01 Jun 2011 15:58:00 -0600 - Pavel Yaskevich
 * Updated to support ALTER (CASSANDRA-1709)
Tue, 22 Mar 2011 18:10:28 -0700 - Eric Evans <>
 * Initial version, 1.0.0