All nodes/Data/DataFrame I/O

Read XLSX

Reads a sheet of an Excel file (xlsx) from the connected file_store into a DataFrame. Supports selecting a sheet by name or index.

Try it

When to use it

Reads one sheet of an Excel workbook into a table (a DataFrame). The store, the path and “Connection source” work exactly as in Read CSV — the difference is the file itself: no delimiter or encoding to guess, but one step reads exactly one sheet.

Use CSV when the file comes from a machine (a dump, an export, an API response), and XLSX when people send you a workbook: formatting, several sheets and dates that CSV would leave you to parse out of a string.

How it works

The “Sheet” field accepts either a name or a number: 0 is the first sheet, Report is the sheet with that name. Blank also means the first sheet. The field supports Jinja, so the sheet name can be assembled from a variable.

“Header row index” says which row of the sheet holds the column names: 0 is the first row. If the table sits under a two-line report title, set 2.

The whole table travels downstream, while the log shows a preview — the first 50 rows, the shape and the column types.

Common mistakes

• A sheet with a numeric name. An all-digits value is always read as an index: a sheet named 2024 cannot be selected this way — the node looks for sheet number 2024 and fails. Rename the sheet or pick it by position.
• The legacy .xls format. Only .xlsx/.xlsm workbooks are read; an old-format file has to be re-saved.
• Expecting every sheet at once. The node returns one sheet. For a second one add a second node and join the results with DF Concat.
• Merged cells in the header. A merged cell’s value lands only in the first of the columns; the rest come back empty, named Unnamed: 2 and so on.
• A workbook over the limit. “Max bytes” (25 MiB) truncates the transfer, and a truncated xlsx cannot be parsed at all: the step fails with a parse error. Such a workbook has to be reduced, or the sheet you need exported as CSV.

Inputs

Outputs

Configuration