DataDriver is a Data-Driven extension for Robot Framework®. This document explains how to use the DataDriver library listener. For information about installation, support, and more, please visit the project page

For more information about Robot Framework®, see https://robotframework.org.

DataDriver is used/imported as Library but does not provide keywords which can be used in a test. DataDriver uses the Listener Interface Version 3 to manipulate the test cases and creates new test cases based on a Data-File that contains the data for Data-Driven Testing. These data file may be .csv , .xls or .xlsx files.

Data Driver is also able to cooperate with Microsoft PICT. An Open Source Windows tool for data combination testing. Pict is able to generate data combinations based on textual model definitions. https://github.com/Microsoft/pict

It is also possible to implement own DataReaders in Python to read your test data from some other sources, like databases or json files.

If you already have Python >= 3.6 with pip installed, you can simply run:

pip install --upgrade robotframework-datadriver

For file support of xls or xlsx file you need to install the extra XLS or the dependencies. It contains the dependencies of pandas, numpy and xlrd. Just add [XLS] to your installation. New since version 3.6.

pip install --upgrade robotframework-datadriver[XLS]

or if you have Python 2 and 3 installed in parallel you may use

pip3 install --upgrade robotframework-datadriver

DataDriver is compatible with Python 2.7 only in Version 0.2.7.

pip install --upgrade robotframework-datadriver==0.2.7

Because Python 2.7 is deprecated, there are no new feature to python 2.7 compatible version.

• What DataDriver Does
• How DataDriver Works
• Usage
• Structure of Test Suite
• Structure of data file
• Accessing Test Data From Robot Variables
• Data Sources
• File Encoding and CSV Dialect
• Custom DataReader Classes
• Selection of Test Cases to Execute
• Configure DataDriver by Pre-Run Keyword
• Pabot and DataDriver

DataDriver is an alternative approach to create Data-Driven Tests with Robot Framework®. DataDriver creates multiple test cases based on a test template and data content of a csv or Excel file. All created tests share the same test sequence (keywords) and differ in the test data. Because these tests are created on runtime only the template has to be specified within the robot test specification and the used data are specified in an external data file.

Brief overview what DataDriver is and how it works at the RoboCon 2020 in Helsiki.

DataDriver gives an alternative to the build in data driven approach like:

This inbuilt approach is fine for a hand full of data and a hand full of test cases. If you have generated or calculated data and specially if you have a variable amount of test case / combinations these robot files become quite a pain. With DataDriver you may write the same test case syntax but only once and deliver the data from en external data file.

One of the rare reasons when Microsoft® Excel or LibreOffice Calc may be used in testing… ;-)

See example test suite

See example csv table

When the DataDriver is used in a test suite it will be activated before the test suite starts. It uses the Listener Interface Version 3 of Robot Framework® to read and modify the test specification objects. After activation it searches for the Test Template -Keyword to analyze the [Arguments] it has. As a second step, it loads the data from the specified data source. Based on the Test Template -Keyword, DataDriver creates as much test cases as data sets are in the data source.

In the case that data source is csv (Default) As values for the arguments of the Test Template -Keyword, DataDriver reads values from the column of the CSV file with the matching name of the [Arguments]. For each line of the CSV data table, one test case will be created. It is also possible to specify test case names, tags and documentation for each test case in the specific test suite related CSV file.

Data Driver is a "Library Listener" but does not provide keywords. Because Data Driver is a listener and a library at the same time it sets itself as a listener when this library is imported into a test suite.

To use it, just use it as Library in your suite. You may use the first argument (option) which may set the file name or path to the data file.

Without any options set, it loads a .csv file which has the same name and path like the test suite .robot .

Example:

In the Moment there are some requirements how a test suite must be structured so that the DataDriver can get all the information it needs.

• only the first test case will be used as a template. All other test cases will be deleted.
• Test cases have to be defined with a Test Template in Settings secion. Reason for this is, that the DataDriver needs to know the names of the test case arguments. Test cases do not have named arguments. Keywords do.
• The keyword which is used as Test Template must be defined within the test suite (in the same *.robot file). If the keyword which is used as Test Template is defined in a Resource the DataDriver has no access to its arguments names.

In this example, the DataDriver is activated by using it as a Library. It is used with default settings. As Test Template the keyword Invalid Login is used. This keyword has two arguments. Argument names are ${username} and ${password}. These names have to be in the CSV file as column header. The test case has two variable names included in its name, which does not have any functionality in Robot Framework®. However, the Data Driver will use the test case name as a template name and replaces the variables with the specific value of the single generated test case. This template test will only be used as a template. The specified data Default and UserData would only be used if no CSV file has been found.

• *** Test Cases *** column has to be the first one.
• Argument columns: For each argument of the Test Template keyword one column must be existing in the data file as data source. The name of this column must match the variable name and syntax.

• [Tags] column may be used to add specific tags to a test case. Tags may be comma separated.
• [Documentation] column may be used to add specific test case documentation.

In this data file, eight test cases are defined. Each line specifies one test case. The first two test cases have specific names. The other six test cases will generate names based on template test cases name with the replacement of variables in this name. The order of columns is irrelevant except the first column, *** Test Cases ***

In general DataDriver supports any Object that is handed over from the DataReader. However the text based readers for csv, excel and so do support different types as well. DataDriver supports Robot Framework® Scalar variables as well as Dictionaries and Lists. It also support python literal evaluations.

The Prefix $ defines that the value in the cell is taken as in Robot Framework® Syntax. String is str, ${1} is int and ${None} is NoneType. The Prefix only defines the value typ. It can also be used to assign a scalar to a dictionary key. See example table: ${user}[id]

Dictionaries can be created in different ways.

One option is, to use the prefix &. If a variable is defined that was (i.e. &{dict}) the cell value is interpreted the same way, the BuiltIn keyword Create Dictionary would do. The arguments here are comma (,) separated. See example table: &{dict}

The other option is to define scalar variables in dictionary syntax like ${user}[name] or ${user.name} That can be also nested dictionaries. DataDriver will create Robot Framework® (DotDict) Dictionaries, that can be accessed with ${user.name.first} See example table: ${user}[name][first]

Lists can be created with the prefix @ as comma (,) separated list. See example table: @{list}

Be aware that a list with an empty string has to be the cell content ${Empty}.

DataDriver can evaluate Literals. It uses the prefix e for that. (i.e. e{list_eval}) For that it uses BuiltIn Evaluate

See example table: e{user.chk}

If neccesary it is possible to access the fetched data tables directly from a Robot Framework® variable. This could be helpfull in Test Setup or in Suite Setup.

There are three variables available within the Data-Driven Suite:

A list as suite variable containing a robot dictionary for each test case that is selected for execution.

This can be accessed as usual in Robot Framework®.

${DataDriver_DATA_LIST}[2][arguments][\\${password}] would result in mode .

A dictionary as suite variable that contains the same data as the list, with the test names as keys.

A dictionary as test variable that contains the test data of the current test case. This dictionary does also contain arguments that are not used in the Test Template keyword. This can be used in Test Setup and within a test case.

By default DataDriver reads csv files. With the Encoding and CSV Dialect settings you may configure which structure your data source has.

To use Excel file types, you have to install DataDriver with the Extra XLS.

If you want to use Excel based data sources, you may just set the file to the extention or you may point to the correct file. If the extention is ".xls" or ".xlsx" DataDriver will interpret it as Excel file. You may select the sheet which will be read by the option sheet_name. By default it is set to 0 which will be the first table sheet. You may use sheet index (0 is first sheet) or sheet name(case sensitive). XLS interpreter will ignore all other options like encoding, delimiters etc.

or:

Microsoft Excel xls or xlsx file have the possibility to type thair data cells. Numbers are typically of the type float. If these data are not explicitly defined as text in Excel, pandas will read it as the type that is has in excel. Because we have to work with strings in Robot Framework® these data are converted to string. This leads to the situation that a European time value like "04.02.2019" (4th January 2019) is handed over to Robot Framework® in Iso time "2019-01-04 00:00:00". This may cause unwanted behavior. To mitigate this risk you should define Excel based files explicitly as text within Excel.

Alternatively you may deactivate that string conversion. To do so, you have to add the option preserve_xls_types to True. In that case, you will get str, float, boolean, int, datetime.time, datetime.datetime and some others.

Pict is able to generate data files based on a model file. https://github.com/Microsoft/pict

Documentation: https://github.com/Microsoft/pict/blob/master/doc/pict.md

• Path to pict.exe must be set in the %PATH% environment variable.
• Data model file has the file extention ".pict"
• Pict model file must be encoded in UTF-8

If the file option is set to a file with the extention pict, DataDriver will hand over this file to pict.exe and let it automatically generates a file with the extention ".pictout". This file will the be used as data source for the test generation. (It is tab seperated and UTF-8 encoded) Except the file option all other options of the library will be ignored.

It is possible to give options to pict with the import argument pict_options=.

This module implements a reader class that creates a test case for each file or folder that matches the given glob pattern.

With an optional argument "arg_name" you can modify the argument that will be set. See folder example.

Example with json files:

Example with folders:

While there are various specifications and implementations for the CSV format (see RFC 4180), there is no formal specification in existence, which allows for a wide variety of interpretations of CSV files. Therefore it is possible to define your own dialect or use predefined. The default is Excel-EU which is a semicolon separated file. These Settings are changeable as options of the Data Driver Library.

• None(default): Data Driver will search in the test suites folder if a *.csv file with the same name than the test suite *.robot file exists
• only file extention: if you just set a file extentions like ".xls" or ".xlsx" DataDriver will search
• absolute path: If an absolute path to a file is set, DataDriver tries to find and open the given data file.
• relative path: If the option does not point to a data file as an absolute path, Data Driver tries to find a data file relative to the folder where the test suite is located.

encoding= must be set if it shall not be cp1252.

Examples:

cp1252, ascii, iso-8859-1, latin-1, utf_8, utf_16, utf_16_be, utf_16_le

cp1252 is:

• Code Page 1252
• Windows-1252
• Windows Western European

Most characters are same between ISO-8859-1 (Latin-1) except for the code points 128-159 (0x80-0x9F). These Characters are available in cp1252 which are not present in Latin-1.

€ ‚ ƒ „ … † ‡ ˆ ‰ Š ‹ Œ Ž ‘ ’ “ ” • – — ˜ ™ š › œ ž Ÿ

See Python Standard Encoding for more encodings

You may change the CSV Dialect here. The dialect option can be one of the following: - Excel-EU - excel - excel-tab - unix - UserDefined

supported Dialects are:

"Excel-EU"
    delimiter=';',
    quotechar='"',
    escapechar='\\',
    doublequote=True,
    skipinitialspace=False,
    lineterminator="\\r\\n",
    quoting=csv.QUOTE_ALL

"excel"
    delimiter = ','
    quotechar = '"'
    doublequote = True
    skipinitialspace = False
    lineterminator = '\\r\\n'
    quoting = QUOTE_MINIMAL

"excel-tab"
    delimiter = '\\t'
    quotechar = '"'
    doublequote = True
    skipinitialspace = False
    lineterminator = '\\r\\n'
    quoting = QUOTE_MINIMAL

"unix"
    delimiter = ','
    quotechar = '"'
    doublequote = True
    skipinitialspace = False
    lineterminator = '\\n'
    quoting = QUOTE_ALL

Usage in Robot Framework®

User may define the format completely free. If an option is not set, the default values are used. To register a userdefined format user have to set the option dialect to UserDefined

Usage in Robot Framework®

file=None,
encoding='cp1252',
dialect='Excel-EU',
delimiter=';',
quotechar='"',
escapechar='\\\\',
doublequote=True,
skipinitialspace=False,
lineterminator='\\r\\n',
sheet_name=0

It is possible to write your own DataReader Class as a plugin for DataDriver. DataReader Classes are called from DataDriver to return a list of TestCaseData.

DataReader classes are loaded dynamically into DataDriver while runtime. DataDriver identifies the DataReader to load by the file extantion of the data file or by the option reader_class.

This will load the class csv_reader from csv_reader.py from the same folder.

This will load the class generic_csv_reader from generic_csv_reader.py from same folder.

Recommendation:

Have a look to the Source Code of existing DataReader like csv_reader.py or generic_csv_reader.py .

To write your own reader, create a class inherited from AbstractReaderClass.

Your class will get all available configs from DataDriver as an object of ReaderConfig on __init__.

DataDriver will call the method get_data_from_source This method should then load your data from your custom source and stores them into list of object of TestCaseData. This List of TestCaseData will be returned to DataDriver.

AbstractReaderClass has also some optional helper methods that may be useful.

You can either place the custom reader with the others in DataDriver folder or anywhere on the disk. In the first case or if your custom reader is in python path just use it like the others by name:

In case it is somewhere on the disk, it is possible to use an absolute or relative path to a custom Reader. Imports of custom readers follow the same rules like importing Robot Framework® libraries. Path can be relative to ${EXECDIR} or to DataDriver/__init__.py:

This my_reader.py should implement a class inherited from AbstractReaderClass that is named my_reader.

See other readers as example.

Because test cases that are created by DataDriver after parsing while execution, it is not possible to use some Robot Framework® methods to select test cases.

Examples for options that have to be used differently:

To execute just a single test case by its exact name it is possible to execute the test suite and set the global variable ${DYNAMICTEST} with the name of the test case to execute as value. Pattern must be suitename.testcasename.

Example:

Pabot uses this feature to execute a single test case when using --testlevelsplit

It is possible to set a list of test case names by using the variable ${DYNAMICTESTS} (plural). This variable must be a string and the list of names must be pipe-seperated (|).

Example:

robot --variable DYNAMICTESTS:firstsuitename.testcase1|firstsuitename.testcase3|anothersuitename.othertestcase foldername

It is also possible to set the variable @{DYNAMICTESTS} as a list variable from i.e. python code.

Because it is not possible to use the command line argument --rerunfailed from robot directly, DataDriver brings a Pre-Run-Modifier that handles this issue.

Normally reexecution of failed testcases has three steps.

• original execution
• re-execution the failed ones based on original execution output
• merging original execution output with re-execution output

The DataDriver.rerunfailed Pre-Run-Modifier removes all passed test cases based on a former output.xml.

Example:

Be aware, that in this case it is not allowed to use ":" as character in the original output file path. If you want to set a full path on windows like e:\\myrobottest\\output.xml you have to use ";" as argument seperator.

Example:

New in 0.3.1

It is possible to use tags to filter the data source. To use this, tags must be assigned to the test cases in data source.

To filter the source, the normal command line arguments of Robot Framework® can be used. See Robot Framework® Userguide for more information Be aware that the filtering of Robot Framework® itself is done before DataDriver is called. This means if the Template test is already filtered out by Robot Framework®, DataDriver can never be called. If you want to use --include the DataDriver TestSuite should have a DefaultTag or ForceTag that fulfills these requirements.

Example: robot --include 1OR2 --exclude foo DataDriven.robot

It is also possible to filter the data source by an init option of DataDriver. If these Options are set, Robot Framework® Filtering will be ignored.

Example:

With config_keyword= it's possible to name a keyword that will be called from Data Driver before it starts the actual processing of the data file. One possible usage is if the data file itself shall be created by another keyword dynamically during the execution of the Data Driver test suite. The config_keyword= can be used to call that keyword and return the updated arguments (e.g. file) back to the Data Driver Library.

The config keyword

• May be defined globally or inside each testsuite individually
• Gets all the arguments, that Data Driver gets from Library import, as a Robot Dictionary
• Shall return the (updated) Data Driver arguments as a Robot Dictionary

Usage in Robot Framework®

You should use Pabot version 1.10.0 or newer.

DataDriver supports --testlevelsplit from pabot only if the PabotLib is in use. Use --pabotlib to enable that.

When using pabot like this, DataDriver automatically splits the amount of test cases into nearly same sized groups. Is uses the processes count from pabot to calculate the groups. When using 8 processes with 100 test cases you will get 8 groups of tests with the size of 12 to 13 tests. These 8 groups are then executed as one block with 8 processes. This reduces a lot of overhead with Suite Setup and Teardown.

You can switch between three modes:

• Equal: means it creates equal sizes groups
• Binary: is more complex. it created a decreasing size of containers to support better balancing.
• Atomic: it does not group tests at all and runs really each test case in a separate thread.

This can be set by optimize_pabot in Library import.

Example:

Binary creates with 40 test cases and 8 threads something like that: