# Data generator properties Properties to configure to be able to generate automatically your test data as a dataset. Data generator is unidirectional and can only be used as a source dataset in your pipeline. ## Data generator connection Select Data generator connection in the list and configure the connection. ## Configuration Select your engine from the list and set the main and advanced settings. After configuring the connection, give it a display name (mandatory) and a description (optional). ## Data generator dataset Dataset configurationDataset configuration | Property | Configuration | | --- | --- | | Dataset name | Enter a
display name for the dataset. This name will be used as a unique identifier of
the dataset in all Talend Cloud apps. | | Connection | Select
your connection in the list. If you are creating a dataset based on an existing
connection, this field is read-only. | | Type | Select the type of dataset you want to create:

Batch if you want to generate records once. The pipeline that uses this dataset will be a batch pipeline.
Streaming if you want to generate records every N milliseconds in a streaming pipeline. The pipeline that uses this dataset will be a streaming pipeline and you will be able to define the polling interval in milliseconds in the Polling configuration field of the source dataset.

| Main settingsMain settings | Property | Configuration | | --- | --- | | Rows | Enter the number of records you want to generate. | | Fields | Define the fields and the nature of the data to be generated. You can
use predefined types in order to help you generate specific data:

Name: Enter the name of the field you want to generate.
Example:
```
firstname 
```
Type: Select in the list the data type you want your field to have. Depending on the type selected, additional fields can be displayed to configure your data.
Example:
```
First Name
```
Blank %: Enter or select the percentage of empty fields you want to generate.
Example: 5, so that five percent of the generated first names are empty fields.

| | Random within list | If you select this type, you can manually add random elements with
custom values and weights to your generated fields.

For example, you can
generate a field named hair\_color with three
elements: brown with 0.4
weight (40% of the generated values), red with
0.4 weight (40% of the generated values) and
blond with 0.2 weight (20% of the generated
values). | Advanced settingsAdvanced settings | Property | Configuration | | --- | --- | | Use seed | Enable this option if you want to use a specific seed to initialize a
random number generator.

Seeds allow you to keep the same
results.

Example: 123456 | | Enable custom locales | Enable this option to or select in the list a custom language and
country code. By default, it is en-us.

You can
select multiple locales: in that case records will be created using the
different selected locales in a random way. It allows you to change the
value of some types according to these locales (for example address
records will vary according to the locale selected). | ## Data generator source dataset configuration Fields to configure in the source dataset of your pipeline. | Property | Configuration | | --- | --- | | Random rows number | Enable this option if you want to generate a random number of rows
according to a minimum (Minimum rows number) and a
maximum (Maximum rows number) number that you
define. | | Polling configuration (only if you created records
using the streaming type) |

On the Main tab, set the time between every generation of a set of records in the Min poll interval field.
On the Advanced tab, set the number of records generated for every set in the Max poll records field. By default, it is 1.