Store and look up data with datasets

A dataset is a small table that belongs to you — columns and rows, like a spreadsheet. A flow can read from it, add to it, change it and do sums on it while a visitor is part way through answering.

Typical uses:

• A price list you look up based on what the visitor chose
• A stock list, so the flow can say what is actually available
• A log of every submission, in your own column layout
• A leaderboard for a quiz

#Create one

Datasets live behind the Datasets button in the Flow Designer header.

Column names are stored in capitals, so Price becomes PRICE. Use that form when you refer to a column later.

Data types available: text, number, email, phone, date and url. The type changes the kind of box you get when editing a cell — and it is worth setting properly, because a number column sorts and totals as a number while a text column does not.

The dataset list shows how many columns and rows each one has. The three-dot menu on a dataset offers Rename and Delete.

#Editing the table

Select a cell to edit it. Enter or Tab saves, Escape cancels. Rows have a delete button; large tables are paged, with Showing 1-50 of 300 under the table.

#Import and export

Import CSV brings a spreadsheet in. You can have Tayon create the columns from the file’s header row, and tell it whether the first row is a header or real data.

Export CSV downloads the table. Do this before any bulk change — it is the only backup available.

#Using a dataset in a flow

Six action types work with datasets. Add an Action step and pick one.

#Recording what a visitor answered

Two actions work as a pair:

The matching is the important bit: an answer stored under the identifier EMAIL goes into the column named EMAIL. If nothing lands in a column, the usual cause is that the identifier and the column name are not the same word.

#Looking something up

Filter DataSet finds the rows that match a condition. You build the condition the same way as a split — a column, a comparison, and a value, which can be a variable such as $REGION. You can also set an Order By and a Max Items limit.

After a filter you can read the results:

The name in square brackets is the dataset’s name.

To print a whole list rather than one row, wrap the layout in a list block. The block repeats once per matching row, and [*] becomes the row number:

[li]
$[Prices].NAME[*] — $[Prices].PRICE[*]
[/li]

#Changing or removing rows

Update DataSet and Delete DataSet both act on whatever the last filter found. So they always come in pairs:

Update DataSet takes the fields to change, and values that can be variables.

#Doing sums

Calculate DataSet works out figures for one column across the filtered rows. The results are then available as:

So a screen can say Your quote comes to $DATASET.SUM after a filter and a calculate.

#Filing one answer straight into a dataset

There is a shortcut for simple cases. On a question’s Interaction tab, Add to Dataset files that answer into a dataset without needing the new/save action pair. Use the actions when you are collecting a whole row across several questions; use this when you just want one answer recorded.

#A pattern that works

A quote calculator, end to end: