Lindorm Cassandra Query Language (CQL) supports JSON in SELECT and INSERT statements. However, JSON support does not fundamentally change the way that you use the Lindorm CQL API. For example, you must define a schema for your database. JSON support only provides a method to work with JSON documents. Lindorm CQL can be used to write or display the data of multiple columns in the JSON format.

SELECT JSON

You can specify the JSON keyword in a SELECT statement. In this case, the SELECT statement returns the data of each row as a JSON-encoded map. The remainder of the behavior for the SELECT statement remains unchanged.

The result map keys are the same as the column names in a regular result set. For example, you can execute the SELECT JSON a, b FROM ... statement to generate a map that contains the a and b keys.

INSERT JSON

You can specify the JSON keyword in an INSERT statement. In this case, the INSERT statement is used to insert a JSON-encoded map as a row. In most cases, the format of the JSON map must match the format of the result set that is returned by a SELECT JSON statement executed on the same table. Take note that column names that are case-sensitive must be enclosed in double quotation marks ("). For example, you can execute the following statement to insert a map into a table that contains the key and value columns: 

INSERT INTO persioninfo JSON '{ "c1": "key", "c2": "value"}'

By default, columns that are unspecified in a JSON map are set to NULL. This rule also applies when the default value NULL is explicitly specified. This means that all existing values of the unspecified columns are removed.

JSON Encoding of CQL Data Types

In some cases, Lindorm CQL represents and accepts data types in the native JSON format. Lindorm CQL also accepts string representations that match the literal format of Lindorm CQL for all single-field types. For example, float values, integers, UUIDs, and dates can be represented by Lindorm CQL literal strings.

The following table describes the data types that Lindorm CQL can accept in INSERT JSON statements and fromJson() functions. The following table also describes the data types that Lindorm CQL can use to return data for SELECT JSON statements and fromJson() functions.
JSON data type Accepted data type Returned data type Description
ascii string string The string for which the \u JSON escaping character is used.
bigint integer, string integer The string must be a valid 64-bit integer.
blob string string The string must be an even number of hexadecimal digits that are prefixed with 0x.
boolean boolean, string boolean Valid values: true and false.
date string string The string that contains a date in the YYYY-MM-DD format. The date is displayed in UTC.
decimal integer, float, string float The precision may exceed that of a 32-bit or 64-bit IEEE-754 floating point number.
double integer, float, string float The string must be a valid integer or a floating-point number.
float integer, float, string float The string must be a valid integer or a floating-point number.
inet string string The IPv4 or IPv6 address.
int integer, string integer The string must be a 32-bit integer.
smallint integer, string integer The string must be a 16-bit integer.
text string string The string for which the \u JSON escaping character is used.
time string string The string that contains a time in the HH-MM-SS[.fffffffff] format.
timestamp integer, string string The timestamp. The string can contain a timestamp in the YYYY-MM-DD HH:MM:SS.SSS format.
timeuuid string string The Type 1 UUID.
tinyint integer, string integer The string must be a valid 8-bit integer.
uuid string string The string in the UUID format.
varint integer, string integer The integer whose length is variable. An integer overflow may occur when a 32-bit or 64-bit integer is transferred to a client-side decoder.