Adding an input
Connect to a file, a database, or a published Tableau data source. The input step then shows every field with checkboxes.
Uncheck fields you do not need immediately. Fewer fields means faster profiling, a faster flow and a smaller output — and it is one click here versus a Remove Field step later.
Wildcard union
On a file input, the Tables tab offers a wildcard union. Point at a folder, give a matching pattern like sales_*.csv, and every matching file is unioned into one input.
Tick Include subfolders where the files are organised by year or month. Prep adds a File Paths field recording where each row came from, which you will want when reconciling.
Input-level filters
The input step can filter rows before anything else runs — by value, by date range, or with a calculation.
This is the cheapest filter available, because excluded rows never enter the flow. Filtering to the last two years at the input rather than at the end can change a flow from minutes to seconds.
Data sample settings
The Data Sample tab controls how much Prep loads while building: a default quick sample, a fixed number of rows, or all data.
Quick sample keeps things responsive on a large source. Switch to all data when you are diagnosing something specific and need every row visible.
Changing field types
The input step shows each field's type and lets you change it before the data enters the flow. Fixing a date stored as text here is better than adding a step later — everything downstream then sees a proper date.