CurrentDatabase¶

Overview¶

The CurrentDatabase expression returns the name of the current database/schema in the Spark session. This is a leaf expression that produces a string value representing the active database context without requiring any input parameters.

Syntax¶

SELECT current_schema();
-- or
SELECT current_database();

// DataFrame API usage
import org.apache.spark.sql.functions._
df.select(expr("current_schema()"))

Arguments¶

Argument	Type	Description
None	N/A	This expression takes no arguments

Return Type¶

Returns a StringType value representing the current database/schema name.

Supported Data Types¶

This expression does not accept input data types as it takes no parameters. It always produces a string output.

Algorithm¶

Retrieves the current database name from the Spark session's catalog context
Returns the active database/schema identifier as a string literal
Operates as a constant expression that doesn't depend on row data
Implements lazy evaluation through the Catalyst expression framework
Marked as Unevaluable, meaning it requires special handling during execution planning

Partitioning Behavior¶

This expression has no impact on partitioning behavior:

Preserves existing partitioning schemes since it's a constant value
Does not require data shuffle or repartitioning
Can be computed independently of data distribution

Edge Cases¶

Never returns null values (nullable = false)
Returns "default" when no specific database is selected
Behavior is consistent across all executors in distributed execution
Thread-safe for concurrent access within the same Spark session

Code Generation¶

This expression is marked as Unevaluable, which means it cannot be directly code-generated using Tungsten. Instead, it requires special handling during query planning where the actual database name is resolved and substituted as a literal value before code generation occurs.

Examples¶

-- Get current database name
SELECT current_schema();
-- Result: default

-- Use in conditional logic
SELECT CASE 
  WHEN current_schema() = 'production' THEN 'prod_mode'
  ELSE 'dev_mode'
END as environment;

// DataFrame API usage
import org.apache.spark.sql.functions._

// Add current database as a column
val df = spark.table("my_table")
  .withColumn("database_name", expr("current_schema()"))

// Filter based on current database
val result = df.filter(expr("current_schema() != 'test_db'"))