Robot Framework Database Library

The Database Library for Robot Framework allows you to query a database and verify the results. It requires an appropriate Python module to be installed separately - depending on your database, like e.g. oracledb or pymysql.

The library consists of some keywords designed to perform different checks on your database. Here you can find the keyword docs.

Wath the talk at Robocon 2024 about the Database Library update.

Requirements

Python
Robot Framework
Python database module you're going to use - e.g. oracledb

Installation

pip install robotframework-databaselibrary

Basic usage examples

*** Settings ***
Library       DatabaseLibrary
Test Setup    Connect To My Oracle DB

*** Keywords ***
Connect To My Oracle DB
    Connect To Database
    ...    oracledb
    ...    db_name=db
    ...    db_user=my_user
    ...    db_password=my_pass
    ...    db_host=127.0.0.1
    ...    db_port=1521

*** Test Cases ***
Get All Names
    ${Rows}=    Query    select FIRST_NAME, LAST_NAME from person
    Should Be Equal    ${Rows}[0][0]    Franz Allan
    Should Be Equal    ${Rows}[0][1]    See
    Should Be Equal    ${Rows}[1][0]    Jerry
    Should Be Equal    ${Rows}[1][1]    Schneider

Person Table Contains Expected Records
    ${sql}=    Catenate    select LAST_NAME from person
    Check Query Result    ${sql}    contains    See
    Check Query Result    ${sql}    equals      Schneider    row=1

Wait Until Table Gets New Record
    ${sql}=    Catenate    select LAST_NAME from person
    Check Row Count    ${sql}    >    2    retry_timeout=5s

Person Table Contains No Joe
    ${sql}=    Catenate    SELECT id FROM person
    ...                    WHERE FIRST_NAME= 'Joe'
    Check Row Count    ${sql}   ==    0

See more examples in the folder tests.

Handling multiple database connections

The library can handle multiple connections to different databases using aliases. An alias is set while creating a connection and can be passed to library keywords in a corresponding argument.

Example

*** Settings ***
Library          DatabaseLibrary
Test Setup       Connect To All Databases
Test Teardown    Disconnect From All Databases

*** Keywords ***
Connect To All Databases
    Connect To Database
    ...    psycopg2
    ...    db_name=db
    ...    db_user=db_user
    ...    db_password=pass
    ...    db_host=127.0.0.1
    ...    db_port=5432
    ...    alias=postgres
    Connect To Database
    ...    pymysql
    ...    db_name=db
    ...    db_user=db_user
    ...    db_password=pass
    ...    db_host=127.0.0.1
    ...    db_port=3306
    ...    alias=mysql

*** Test Cases ***
Using Aliases
    ${names}=    Query    select LAST_NAME from person    alias=postgres
    Execute Sql String    drop table XYZ                  alias=mysql

Switching Default Alias
    Switch Database    postgres
    ${names}=    Query    select LAST_NAME from person
    Switch Database    mysql
    Execute Sql String    drop table XYZ

Using configuration file

The Connect To Database keyword allows providing the connection parameters in two ways:

As keyword arguments
In a configuration file - a simple list of key=value pairs, set inside an alias section.

You can use only one way or you can combine them:

The keyword arguments are taken by default
If no keyword argument is provided, a parameter value is searched in the config file

Along with commonly used connection parameters, named exactly as keyword arguments, a config file can contain any other DB module specific parameters as key/value pairs. If same custom parameter is provided both as a keyword argument and in config file, the keyword argument value takes precedence.

The path to the config file is set by default to ./resources/db.cfg. You can change it using an according parameter in the Connect To Database keyword.

A config file must contain at least one section name - the connection alias, if used (see Handling multiple database connections), or [default] if no aliases are used.

Config file examples

Config file with default alias (equal to using no aliases at all)

[default]
db_module=psycopg2
db_name=yourdbname
db_user=yourusername
db_password=yourpassword
db_host=yourhost
db_port=yourport

Config file with a specific alias

[myoracle]
db_module=oracledb
db_name=yourdbname
db_user=yourusername
db_password=yourpassword
db_host=yourhost
db_port=yourport

Config file with some params only

[default]
db_password=mysecret

Config file with some custom DB module specific params

[default]
my_custom_param=value

Inline assertions

Keywords, that accept arguments assertion_operator and expected_value, perform a check according to the specified condition - using the Assertion Engine.

Examples

Check Row Count     SELECT id FROM person          ==        2
Check Query Result  SELECT first_name FROM person  contains  Allan

Retry mechanism

Assertion keywords, that accept arguments retry_timeout and retry_pause, support waiting for assertion to pass.

Setting the retry_timeout argument enables the mechanism - in this case the SQL request and the assertion are executed in a loop, until the assertion is passed or the retry_timeout is reached. The pause between the loop iterations is set using the retry_pause argument.

The argument values are set in Robot Framework time format - e.g. 5 seconds.

The retry mechanism is disabled by default - retry_timeout is set to 0.

Examples

${sql}=   Catenate    SELECT first_name FROM person
Check Row Count     ${sql}  ==        2      retry_timeout=10 seconds
Check Query Result  ${sql}  contains  Allan  retry_timeout=5s  retry_pause=1s

Logging query results

Keywords, that fetch results of a SQL query, print the result rows as a table in RF log.

A log head limit of 50 rows is applied, other table rows are truncated in the log message.
The limit and the logging in general can be adjusted any time in your tests using the Keyword Set Logging Query Results.

You can also setup the limit or disable the logging during the library import.

Examples

*** Settings ***
# Default behavior - logging of query results is enabled, log head is 50 rows.
Library    DatabaseLibrary

# Logging of query results is disabled, log head is 50 rows (default).
Library    DatabaseLibrary    log_query_results=False

# Logging of query results is enabled (default), log head is 10 rows.
Library    DatabaseLibrary    log_query_results_head=10

# Logging of query results is enabled (default), log head limit is disabled (log all rows).
Library    DatabaseLibrary    log_query_results_head=0

Commit behavior

While creating a database connection, the library doesn't explicitly set the autocommit behavior - so the default value of the Python DB module is used. According to Python DB API specification it should be disabled by default - which means each SQL transaction must contain a dedicated commit statement, if necessary.

The library manages it for you:

Keywords like Execute SQL String perform automatically a commit after running the query - or a rollback in case of error
Keywords like Query don't perform a commit, but also do a rollback in case of error

You can turn off this automatic commit/rollback behavior using the no_transaction parameter. See docs of a particular keyword.

It's also possible to explicitly set the autocommit behavior on the Python DB module level - using the Set Auto Commit keyword. This has no impact on the automatic commit/rollback behavior in library keywords (described above).

Database modules compatibility

The library is basically compatible with any Python Database API Specification 2.0 module.

However, the actual implementation in existing Python modules is sometimes quite different, which requires custom handling in the library. Therefore there are some modules, which are "natively" supported in the library - and others, which may work and may not.

Python modules currently "natively" supported

Oracle

oracledb
- Both thick and thin client modes are supported - you can select one using the oracle_driver_mode parameter.
- However, due to current limitations of the oracledb module, it's not possible to switch between thick and thin modes during a test execution session - even in different suites.
cx_Oracle