Data Preparation and Input/Output
Inputs, outputs, and field metadata
Almost every workflow starts at the Input Data tool. Its Options tab lets you override the file format, flag that the first row contains field names, choose the delimiter, set a code page (character encoding), skip junk header rows with Start Data on Line, and cap rows with a Record Limit while testing. The Fields tab shows the metadata Alteryx inferred - name, type, size - and lets you correct it before anything runs downstream. Learn the types: V_String (variable-length text, 254 characters by default), String/WString (fixed length), Int16/32/64, Float, Double, Fixed Decimal, Date, DateTime, Bool, Blob, and SpatialObj.
The Directory tool lists files in a folder, returning fields such as FullPath, FileName, Directory, and Size; it pairs with Dynamic Input (a template file plus a field holding full paths) to read an entire folder through one tool. The Output Data tool writes results, offering overwrite vs. append, code page selection, Max Records per File splitting, and Take File/Table Name From Field to write many files from one stream.
Traps
.yxdb is fastest and preserves types; CSV and Excel are text and lose type information. Text values longer than the declared field size are truncated on read, so watch the field-size warning. A wrong code page produces garbled characters, and numeric-looking text can import as a string.
Rule of thumb
If a CSV has three title rows above the header, set Start Data on Line to 4. Profile and fix types right after the Input tool, never mid-stream.