Skip to main content
Version: 0.17.6


Below are some common questions and their answers with regards to working with your data in DataChat. If your questions aren't answered here, please use the search bar or the sidebar to look through the rest of the documentation. You can also contact us.

How do I load data from a file?

You can load data from comma-separated (.csv) files and Excel (.xls and .xlsx) files. You can even upload zipped (.zip) files that contain .csv and .xls/.xlsx files. To load these files, use Load data from the file <file name> utterance, where <file name> is the name of the file you want to upload.

How do I load data from a directory?

If the files you want to load into DataChat live on a network directory, you can use the utterance Load data from the directory <directory name>, where <directory name> is the network location of the directory.

How do I see all of the datasets I've loaded?

Use the List all datasets utterance to see all of the datasets in your workflow.

How do I use autocomplete?

Whenever you enter a space, DataChat's autocomplete functionality attempts to help you complete your utterance by suggesting values or variables. You can use the arrow keys to select one of the suggestions. Then, use Enter to select that suggestion and add it to your utterance. You can also click the suggestion you want or simply continue typing.

Autocomplete doesn't show me an option that I think it should. What should I do?

There are a few reasons why autocomplete isn't showing you an option you think it should:

  • The list of suggestions is long. Autocomplete only shows the first few suggestions. You can use the arrow keys or your mouse to scroll the list of suggestions and see if the option you're looking for is farther down the list.
  • Autocomplete can't make a suggestion. For example, if you're looking for matches based on a numeric value, you would simply enter the value you're looking for.
  • Finding all of the suggestions is too expensive. If you have a very large dataset, finding all of the potential suggestions or values would take too long to find. In this case, autocomplete won't show any suggestions. Instead, you can browse the underlying dataset using an utterance like Show the first 1000 rows followed by Show the next 1000 rows to find the value you're looking for. After you find that value, you can go back to your original utterance and use that value. Note that if you're looking for a string value, you can use the Contains command to look for words or phrases anywhere in a string.

The text in the chat box is gray. What does that mean?

Gray text means that your current utterance is not a complete sentence in DataChat English. You'll need to add additional phrases (with the help of autocomplete) to finish the sentence. The text will turn black when the sentence is complete.

How do I combine two datasets?

The Extend and Concatenate skills are the main ways to combine datasets.

  • Use Extend if you want to combine datasets based on a common column (also known as a "join"). For example, if you had a Pet dataset with an OwnerID column and an Owner dataset with the same OwnerID column, you would use the utterance Extend the dataset called Pets with the dataset called Owners. This creates a single dataset containing all of the rows between the Pets and Owners datasets that match on that common OwnerID column.
  • Use Concatenate if you need to add rows from one dataset as new rows in another. For example, if you had two datasets, each with 50 rows, the Concatenate skill will create a single dataset with 100 rows. DataChat can also automatically remove duplicate rows for you, if needed. Note that you can only concatenate two datasets that have the same columns because the rows must all have the same format.

How do I add a column to a dataset?

Use the Create skill to create a new column and add it to your current dataset with the utterance Create a new column called <column name>. Note that the Create skill has a few options you can try to add values to your new column:

  • You can concatenate a string from other string columns by saying, for example, Create a new column called <column name> by concatenating <column 1>, -, <column 2>. This takes the value from columns 1 and 2 and concatenates them together using a dash (-).
  • You can calculate the value of your new column using the values of other columns in a math expression by saying, for example, Create a new column called <column name> as (<column 1> + <column 2>) * <column 3> / <column 4>. This then populates your new column with the result of that expression.
  • You can create an index column by saying, for example, Create a new column called <column name> as an increasing sequence.
  • You can create a string column containing a specific string value by saying, for example, Create a new column called <column name> with text Example Text. This creates a new column and fills each row of the column with "Example Text."

How do I name a dataset?

Use the Rename skill to rename an existing dataset. For example, if you wanted to rename the current dataset to "MarketingData," you would say Rename the current dataset MarketingData.

There are too many datasets in my view. How can I clean them up?

Use the Forget skill to remove a dataset. Note that the dataset no longer shows up when using the List skill or in any autocomplete suggestions. However, forgetting a dataset does not delete the uploaded file it's based on.

I forgot what I named a dataset. How can I find that dataset?

Use the List all datasets utterance to see all of the datasets in your workflow. Alternatively, you can use the Toggle Explorer option in the menu to see all of the utterances in your workflow so far, which can help you remember what you named a dataset.

How can I leave a note in my workflow?

Use the Record a <color> note <note contents> utterance to leave a note. If no color is specified, a white note is recorded.

How do I save my work?

There are several options for saving your work in DataChat:

  • If you want to save your final dataset, use the Export command to save your dataset as a .csv file. For example, you can say Export my current dataset to <file name>, where <file name> is the name of your new .csv file.
  • If you want to save a chart or a plot, you can click the download button or right-click on it and export it as an image.
  • If you want to save your steps from this session (i.e. your workflow), use the Save this workflow as <workflow name> utterance. After a workflow has been saved, you can automatically run each step in the workflow again by saying Replay the workflow <workflow name>.

How do I share my work?

There are several options for sharing your work in DataChat:

  • You can create an Insights Board to show your final results. Insight Boards let you present your data in a flexible grid format rather than the typical notebook format. You can share your Insights Board with other users at your organization by using the Share menu. For more information on creating an Insights Board, refer to the Create an Insights Board topic.
  • You can use the Share skill to collaborate on a saved workflow with other users at your organization. Note that sharing a workflow does not automatically grant access to the files or databases referenced in the workflow.
  • You can use the Print Magazine option from the menu within your app to download a PDF of all of your results that you can share with other users.