Exasol MCP Server Tools#

Tools Providing Information About Custom Database Objects #

list_exasol_schemas #

Description:

Lists database schemas. Visibility of schemas can be restricted in the settings. Can be hidden globally via enable_list_tools in the settings (see Tool Setup).

Returns:

Type: list
Data:
- name: name of the schema
- comment: schema comment, if available

Finds database schemas by looking for the specified keywords in their names and comments. Visibility of schemas can be restricted in the settings. Can be hidden globally via enable_find_tools in the settings (see Tool Setup).

Returns:

Type: list
Data:
- name: name of the schema
- comment: schema comment, if available

list_exasol_tables_and_views #

Description:

Lists tables and views in the specified database schema. Visibility of tables and views can be restricted in the settings. Can be hidden globally via enable_list_tools in the settings (see Tool Setup).

Returns:

Type: list
Data:
- schema: name of the schema where the table or view is located
- name: name of the table or view
- comment: table or view comment, if available

find_exasol_tables_and_views #

Description:

Finds tables and views by looking for the specified keywords in their names and comments. Optionally, limits the search to one specified schema. Visibility of tables and views can be restricted in the settings. Can be hidden globally via enable_find_tools in the settings (see Tool Setup).

Returns:

Type: list
Data:
- schema: name of the schema where the table or view is located
- name: name of the table or view
- comment: table or view comment, if available

list_exasol_custom_functions #

Description:

Lists custom functions in the specified database schema. Visibility of functions can be restricted in the settings. Can be hidden globally via enable_list_tools in the settings (see Tool Setup).

Returns:

Type: list
Data:
- schema: name of the schema where the function is located
- name: name of the function
- comment: function comment, if available

find_exasol_custom_functions #

Description:

Finds custom functions by looking for the specified keywords in their names and comments. Optionally, limits the search to one specified schema. Visibility of functions can be restricted in the settings. Can be hidden globally via enable_find_tools in the settings (see Tool Setup).

Returns:

Type: list
Data:
- schema: name of the schema where the function is located
- name: name of the function
- comment: function comment, if available

list_exasol_user_defined_functions #

Description:

Lists User Defined Functions (UDF) in the specified database schema. Visibility of UDFs can be restricted in the settings. Can be hidden globally via enable_list_tools in the settings (see Tool Setup).

Returns:

Type: list
Data:
- schema: name of the schema where the UDF is located
- name: name of the UDF
- comment: UDF comment, if available

find_exasol_user_defined_functions #

Description:

Finds User Defined Functions (UDF) by looking for the specified keywords in their names and comments. Optionally, limits the search to one specified schema. Visibility of UDFs can be restricted in the settings. Can be hidden globally via enable_find_tools in the settings (see Tool Setup).

Returns:

Type: list
Data:
- schema: name of the schema where the UDF is located
- name: name of the UDF
- comment: UDF comment, if available

list_exasol_preprocessors #

Description:

Lists available SQL preprocessor scripts and reports which one is currently active in the session. Can be disabled in the settings (see Tool Setup).

Returns:

Type: dict
Data:
- preprocessors: list of preprocessor scripts, each with:
  
  schema: name of the schema where the preprocessor is located
  
  name: name of the preprocessor script
  
  comment: script comment, if available
- current_preprocessor: fully-qualified name of the active preprocessor (e.g. MY_SCHEMA.MY_PREPROCESSOR), or null if none is set

set_exasol_preprocessor #

Description:

Activates a SQL preprocessor script at the session level. This setting is not persistent: the MCP server manages the database connection independently and may reconnect without notice, silently resetting the active preprocessor. Before running queries that depend on a preprocessor, the agent should verify the active setting with list_exasol_preprocessors and re-apply if necessary. Can be disabled in the settings (see Tool Setup).

Arguments:

schema_name: name of the schema containing the preprocessor script
script_name: name of the preprocessor script

Returns:

Type: string
Data: fully-qualified name of the newly-activated preprocessor

describe_exasol_tables_and_views #

Description:

Provides full available information about one or more specified tables and/or views. Pass the names of every table or view needed in a single call.

Parameters:

schema_name: name of the schema containing the tables and/or views
table_names: list of table and/or view names to describe

Returns:

Type: list
Data: one entry per requested name, in the order requested:
- schema: name of the schema where the table or view is located
- name: name of the table or view
- comment: table or view comment, if available
- columns: list of table or view columns
  
  name: column name
  
  comment: column comment, if available
  
  type: SQL type, e.g. “VARCHAR(2000)”
- constraints: for tables only, list of table constraints if there are any:
  
  name: constraint name
  
  constraint_type: constraint type - either “PRIMARY KEY” or “FOREIGN_KEY”
  
  columns: comma separated list of columns the constraint is applied to
  
  referenced_schema: schema referenced in the FOREIGN KEY constraint
  
  referenced_table: table referenced in the FOREIGN KEY constraint
  
  referenced_columns: comma separated list of columns in the referenced table in the FOREIGN KEY constraint

If any requested name is not found, the whole call fails with an error naming the missing table(s) or view(s).

summarize_exasol_table #

Description:

Summarizes the content of a table or view. Returns the total row count and a configurable number of sample rows (default: 10). For each column returns:

most common distinct values in descending frequency order
number of distinct non-NULL values
presence of NULL values and their percentage
minimum and maximum values (numeric columns only)

Must be explicitly enabled in the settings (see Tool Setup).

Parameters:

schema_name: name of the schema
table_name: name of the table or view
sample_size (optional, default 10): number of sample rows to include, between 1 and 100
top_values (optional, default 5): number of most common distinct values to return per column, between 1 and 100

Returns:

Type: dict
Data:
- schema: name of the schema where the table or view is located
- name: name of the table or view
- comment: table or view comment, if available
- row_count: total number of rows in the table or view
- columns: list of column statistics, each column contains:
  
  name: column name
  
  comment: column comment, if available
  
  type: SQL type, e.g. “DECIMAL(18,0)”
  
  distinct_count: number of distinct non-NULL values
  
  min: minimum value for numeric columns, null otherwise
  
  max: maximum value for numeric columns, null otherwise
  
  top_values: most common distinct values in descending frequency order; empty list if all values are NULL
  
  has_nulls: true if the column contains at least one NULL value
  
  null_percentage: percentage of NULL values rounded to whole percent
- sample: sample rows, in the shape controlled by the query_result_format setting (see Tool Setup) - by default {"columns": [...], "rows": [[...], ...]}, or a list of dicts with column names as keys if query_result_format is set to "dict"

describe_exasol_custom_functions #

Description:

Provides full available information about one or more specified custom functions. Pass the names of every function needed in a single call.

Parameters:

schema_name: name of the schema containing the functions
func_names: list of function names to describe

Returns:

Type: list
Data: one entry per requested name, in the order requested:
- input: list of input parameters
  
  name: parameter name
  
  type: SQL type, e.g. “VARCHAR(2000)”
- returns: returned SQL type

If any requested name is not found, the whole call fails with an error naming the missing function(s).

describe_exasol_user_defined_functions #

Description:

Provides full available information about one or more specified User Defined Functions (UDFs). Pass the names of every UDF needed in a single call.

Parameters:

schema_name: name of the schema containing the UDFs
func_names: list of UDF names to describe

Returns:

Type: list
Data: one entry per requested name, in the order requested:
- input: list of input parameters
  
  name: parameter name
  
  type: SQL type, e.g. “VARCHAR(2000)”
- dynamic_input: indication that the UDF accepts dynamic input
- returns: for return type UDF, returned SQL type
- emits: for emit type UDF, the list of output columns:
  
  name: parameter name
  
  type: SQL type
- dynamic_output: for emit type UDF, indication that the UDF emits dynamic output

If any requested name is not found or fails to parse, the whole call fails with an error naming the problem UDF(s).

Tools Executing a Query #

execute_exasol_query #

Description:

Executes the specified query, which must be a SELECT statement. The query should not modify the data. SELECT INTO command is not allowed. An optional row_limit parameter can be used to preview a sample of results without fetching all rows. The query is then wrapped in SELECT * FROM (<query>) LIMIT <row_limit>. If row_limit is omitted, the server’s configured default_row_limit setting (see Tool Setup), if any, is applied instead.

Returns:

Type: dict or list, depending on the query_result_format setting (see Tool Setup)
Data:
- by default (query_result_format = "tabular"): a dict with columns (list of column names) and rows (list of rows, each row a list of values in the same order as columns)
- if query_result_format is set to "dict": selected rows in a form of dictionaries, with column names as keys

profile_exasol_query #

Description:

Runs the specified SELECT query with profiling enabled and returns a breakdown of the execution plan. Use this to understand why a query is slow.

Returns:

Type: dict or list, depending on the query_result_format setting (see Tool Setup)
Data: the execution plan breakdown, in the shape controlled by query_result_format (see execute_exasol_query above). Each row has the following columns:
- PART_NAME: name of the execution step
- PART_INFO: additional information about the step
- OBJECT_SCHEMA: schema of the database object involved
- OBJECT_NAME: name of the database object involved
- OBJECT_ROWS: number of rows processed
- DURATION: duration of the step in milliseconds
- CPU: CPU time in milliseconds

execute_exasol_write_query #

Description:

Executes the specified DML or DDL query. Normally, this tool can be used only if the MCP Client supports elicitation. The user must review and approve the query execution. The elicitation also allows altering the query.

Returns:

Type: str or None
Data:
None if the query was executed it its original form, otherwise modified query

Tools for Reading, Writing and Deleting Files in BucketFS #

list_bucketfs_directories #

Description:

Lists subdirectories of the specified BucketFS directory.

Returns:

Type: list
Data:
subdirectory paths relative to the bucket root or the root path provided in the settings

list_bucketfs_files #

Description:

Lists files in the specified BucketFS directory.

Returns:

Type: list
Data:
file paths relative to the bucket root or the root path provided in the settings

find_bucketfs_files #

Description:

Finds files in the specified BucketFS directory by looking for the provided keywords in their paths. Files are searched in the given directory and all its descendant subdirectories.

Returns:

Type: list
Data:
file paths relative to the bucket root or the root path provided in the settings

read_bucketfs_text_file #

Description:

Reads the content of a BucketFS text file.

Returns:

Type: str
Data:
file content

write_text_to_bucketfs_file #

Description:

Writes the provided text to a file in BucketFS. The new file will overwrite an existing file. Normally, this tool can be used only if the MCP Client supports elicitation. The user must confirm the operation via elicitation.

Returns:

Type: None

download_file #

Description:

Downloads a file from a given url and saves it at the specified path in BucketFS. The new file will overwrite an existing file. Normally, this tool can be used only if the MCP Client supports elicitation. The user must confirm the operation via elicitation.

Returns:

Type: None

delete_bucketfs_file #

Description:

Deletes BucketFS file at the specified path. Normally, this tool can be used only if the MCP Client supports elicitation. The user must confirm the operation via elicitation.

Returns:

Type: None

delete_bucketfs_directory #

Description:

Deletes a BucketFS directory at the specified path. This operation will recursively delete all files and all subdirectories in this directory. Normally, this tool can be used only if the MCP Client supports elicitation. The user must confirm the operation via elicitation.

Returns:

Type: None

Tools Providing Information About Exasol SQL Dialect #

list_exasol_sql_types #

Description:

Lists Exasol SQL types and their parameters. Can be hidden globally via enable_dialect_tools in the settings (see Tool Setup).

Returns:

Type: list
Data:
- type: SQL type, e.g. “DECIMAL(10,5)”
- create_params: type parameters to be specified when creating a table column
- precision: default precision where applicable

list_exasol_system_tables #

Description:

Lists Exasol system tables in the SYS schema. Can be hidden globally via enable_dialect_tools in the settings (see Tool Setup).

Returns:

Type: list
Data:
- table name

describe_exasol_system_table #

Description:

Provides full available information about the specified system table. Can be hidden globally via enable_dialect_tools in the settings (see Tool Setup).

Returns:

Type: dict
Data:
- schema: name of the schema - SYS
- name: name of the system table
- comment: table comment

list_exasol_statistics_tables #

Description:

Lists Exasol statistics tables in the EXA_STATISTICS schema. Can be hidden globally via enable_dialect_tools in the settings (see Tool Setup).

Returns:

Type: list
Data:
- table name

describe_exasol_statistics_table #

Description:

Provides full available information about the specified statistics table. Can be hidden globally via enable_dialect_tools in the settings (see Tool Setup).

Returns:

Type: dict
Data:
- schema: name of the schema - EXA_STATISTICS
- name: name of the statistics table
- comment: table comment

list_exasol_keywords #

Description:

Lists Exasol keywords that start with a given letter. A tool argument specifies if the output should include keywords that are reserved words or not reserved words. Can be hidden globally via enable_dialect_tools in the settings (see Tool Setup).

Returns:

Type: list
Data:
- keyword

list_exasol_built_in_function_categories #

Description:

Lists built-in function categories. Can be hidden globally via enable_dialect_tools in the settings (see Tool Setup).

Returns:

Type: list
Data:
- function category

list_exasol_built_in_functions #

Description:

Lists built-in functions in the specified category. Can be hidden globally via enable_dialect_tools in the settings (see Tool Setup).

Returns:

Type: list
Data:
- function name

describe_exasol_built_in_function #

Description:

Provides full available information about the specified built-in function. Returns a list that in most cases includes one data structure. However, in few cases, when a function with the same name works with different types of data, the tool returns multiple structures. Can be hidden globally via enable_dialect_tools in the settings (see Tool Setup).

Returns:

Type: list
Data:
- name: function name
- alias: alternative function name, if available
- types: comma-separated list of categories the function belongs to
- description: description of data returned by the function
- purpose: more detailed description of the function, if available
- syntax: call syntax, if available
- usage: guidelines, restrictions and limitations, if applicable
- example: one or more call examples

Error Handling #

If a database-backed tool (e.g. execute_exasol_query) fails because of an error returned by the Exasol server or the underlying driver, the error message returned to the client never includes internal connection details such as the DSN, database or OS username, OS name, or driver/client version. For errors specific to the query itself (for example a SQL syntax error or a reference to a non-existent object) the message includes the original database error text, since that text does not carry any such internal information. For all other database-level errors (for example connection or authentication failures) a generic message is returned instead; full details are only available in the server-side log.