[Yandex Cloud documentation](../../index.md) > [Yandex DataLens](../index.md) > Troubleshooting > [FAQ](index.md) > Datasets

# Datasets

### How do I change data in a source using DataLens? {#how-to-change-source-data}

You cannot change data in a source using DataLens.

You can process data you get from a source on the DataLens side using [calculated fields](../concepts/calculations/index.md).


### The date field is recognized as a string. What should I do? {#incorrectly-identified-date-field}

DataLens uses dates in ISO format. This means, if a date in your source data is `01.01.2020` (its format is `DD.MM.YYYY`), DataLens will treat it as a string.

To convert this kind of field value to the `Date` type, create a new field with this formula: `DATE_PARSE([field_name_with_date_in_DD.MM.YYYY_format])`.

The [`DATETIME_PARSE`](../function-ref/DATETIME_PARSE.md) function works similarly and converts a string to the `Date and Time` type.

### How do I link two tables with the JOIN operator if the fields to link have different data types? {#how-to-link-two-tables}

Fields to link your tables by cannot have different data types.
To link tables, convert the fields in the source table to the type you need.

### How do I add row numbering to a table? {#add-line-numeration}

To add row numbering, use a calculated field, e.g., `RSUM(MIN(1))`. See the [screenshot](https://storage.yandexcloud.net/doc-files/add-line-numeration.png).

### How do I convert a Unix time field to the DataLens Date and time field? {#unix-time}

To do this, convert the Unix time to seconds and use the [DATETIME](../function-ref/DATETIME.md) function. For example, convert this Unix time in milliseconds: `DATETIME(1380717142785/1000)`, where the `/1000` operation converts milliseconds to seconds.

For Unix time fields, the data type in DataLens is an integer or fraction.

### Can I use SQL queries to generate a dataset?

You can directly query a database via the dataset creation interface using [SQL queries](../dataset/settings.md#sql-request-in-datatset).

### How can I get a point's coordinates from its address? {#make-geocode}

You could use the `GEOCODE()` function for that; however, it is not supported anymore.
Use the Yandex Maps API [geocoder](https://yandex.com/dev/maps/geocoder) instead. Please note the API [license terms](https://yandex.com/dev/maps/commercial/doc/concepts/jsapi-geocoder.html).

See [this example of a Jupyter notebook](https://github.com/yandex-cloud-examples/yc-datalens-get-coordinates-by-location) with address geocoding scripts.

### Where can I get geopolygons/geopoints for regions/districts/cities? {#get-geo}

You can use ready-made [geodata sets](https://geointellect.com/files/geo_for_datalens.zip) in DataLens format from Geointellect, our partner.

The archive contains the following data:

* Countries (polygons and points).
* Russian regions (polygons and points).
* Russian cities (points).
* Districts of million-plus cities (polygons).

### Why is dataset materialization not available? {#where-is-materialization}

Materialization is no longer supported. We recommend working with a database directly. If a database responds to DataLens analytical queries too slowly, build data marts, e.g., based on [Yandex Managed Service for ClickHouse®](../../managed-clickhouse/index.md). You can use [Yandex Data Transfer](../../data-transfer/index.md) to upload your data.



_ClickHouse® is a registered trademark of [ClickHouse, Inc](https://clickhouse.com)._