Skip to content

Import Data

Pull a previously uploaded file into a workflow.

When to use this

  • You want to feed an uploaded CSV or Excel file into a workflow
  • Every workflow that starts from raw data begins with one or more Import Data nodes

What you need

  • A data source holding a CSV or Excel file.

How it works

Open the Import Data node and pick a data source from the Data Source selector in the side panel. File says which of its files the node reads:

  • Latest file, the default, reads the data source's newest file. When a newer file is added, such as next month's export, the next run reads it without the workflow being edited.
  • A file chosen by name reads that one file, whatever is added later. Choose one to reproduce an earlier analysis.

The node shows the columns and row count of the file it reads, and emits one row per file row to whatever you wire downstream. That is typically a mapping node from the installed package or an operation node (Join, Filter, Aggregate, Script).

When a run is accepted, the node's choice is resolved to one file, and the run records that file and keeps it fixed for that run. A file added while the run waits is the next run's. If the data source has no file, or no longer holds the chosen one, the run is refused and says which.

An Excel workbook is read as one table: its first sheet, unless that sheet holds no table and exactly one other sheet does. Any other sheet that holds values is not read, and a warning on the upload names it. To use a table on another sheet, move that sheet to the front of the workbook, or save it as CSV, and upload again.

Every cell arrives downstream as text, numbers and dates included: a quantity of 12 reaches the next node as "12". A whole number from a workbook loses its decimal point, so 1001.0 arrives as "1001", and an empty cell arrives as empty text, "". Filter compares numbers as numbers when both sides read as numbers, and a mapping node converts each column to the type its field needs. In a Script, convert with float(...) or int(...) before calculating.

Column headers come straight from the file, with their original capitalization preserved. A header the file repeats (ignoring surrounding spaces) is read with a number added: the second cost becomes cost_2, and a Column headers repeated warning names the renamed column.

Workflow wiring

flowchart LR
    n1["Import Data (records, latest file)"] --> n2["Filter"]
    n2 --> n3["Save Data"]

Use one Import Data per data source. To assemble a scenario, pass the rows through the installed package's mapping nodes and connect their typed outputs to Create Scenario.

Common mistakes

  • Selecting the wrong data source in the Data Source picker. Open the node and pick again
  • Choosing a file by name and then expecting the next month's file to be read. Set File back to Latest file
  • Forgetting that column names keep the file's original capitalization. Filter and Aggregate match column names exactly
  • Keeping the table on the second sheet of a workbook. Only one sheet is read; move the sheet you need to the front
  • Adding numbers in a Script without converting them first. "12" + "3" is "123", not 15

Messages this node can show

Like every node, it can also stop on the checks before a run and the run's limits.