CSV column mapper

A CSV column mapper changes the structure of a table while keeping each record's values associated with their source fields. You can rename headers, change their order, omit unnecessary columns, and add columns containing a fixed value. This is useful when a destination expects SKU, Name, and Region but a supplier sends item_code, description, region, and an internal note. The mapper makes the correspondence explicit instead of relying on a position that happens to look correct.

Import and confirm the source table before editing the mapping. An incorrect delimiter can make several intended fields appear as one, and an incorrect header selection can treat the first record as a column name. Resolve empty or duplicate source headers during parsing. Once the source structure is clear, the mapper presents an ordered list of destination columns, each linked to a source field or an explicit constant. You remain responsible for knowing what the destination system expects.

For a source-linked destination column, choose its source and enter the target name. Renaming item_code to SKU changes the output header while values such as 001 remain the same strings. A new name does not convert numbers, infer dates, trim whitespace, or change the meaning of the values. If a target column requires cleaned descriptions, apply a separate cleaning step and inspect those transformations before mapping the final structure.

Target names must be nonempty and unique. Two columns both named Status would be ambiguous to a receiving program, so the mapper blocks that output. Case differences are preserved as written; you should follow the destination's exact naming requirements rather than assuming it will ignore case. A name containing punctuation can still be valid CSV, but it may be unacceptable to a specific application. The workbench checks its own structural rules and does not claim universal target-system compatibility.

Use the up and down controls to arrange destination fields in the required order. The entire value column moves with its header, so changing the order does not shift one row independently from another. A removed mapping means that source field is absent from the output structure. The original table is retained in the session, and the mapper reports the number of source columns omitted. Review the final header sequence before exporting a file intended for a positional importer.

Internally, selected source columns use stable identifiers. A previous rename or reorder does not by itself make a later rule point to an unrelated field merely because that field now occupies the old position. If a required source column actually disappears, the workflow blocks the invalid reference and asks for a new mapping. This is safer than treating the third visible column as permanently equivalent to the third column from the original file.

A constant column repeats exactly the supplied value in every output record. It can provide a known batch label, a destination flag, or a source-independent category you deliberately assign. A constant is not a calculation or a lookup, and the tool does not infer it from other cells. Empty constants are allowed when the destination expects a blank field, but they do not demonstrate that required business information has been supplied.

The required-columns input lists destination header names separated by commas. If a listed name is missing from the mapping, the operation is blocked. This verifies the presence of a column, not that every cell in it is populated or valid. The mapper also requires at least one actual source-linked column so a wholly disconnected constant table cannot masquerade as a successful mapping of the uploaded data. More elaborate data validation belongs to a separate, explicitly defined task.

Consider a source with item_code, description, region, and amount. Its first row contains 001, a product description, North, and 10. Map item_code to SKU, description to Name, and region to Region, then arrange them in that order. Omit amount if the destination does not need it. Add a constant column named Batch with value September. The first output row contains the original code and description, the original region, and the deliberate batch label.

Mark SKU, Name, and Region as required column names, then run and inspect the result. The presence check should pass because those destination headers exist. If a particular record has an empty region, that cell remains empty; the mapper does not claim that all business requirements are satisfied. You can now apply the result and split by Region or by record count. Inspect any empty-region group separately before deciding whether the receiving system can accept it.

Save rules downloads the current tool configuration together with the source header structure. It does not include the table's records. Load rules checks the incoming table's columns before applying a saved configuration. A mismatched structure is rejected so you can remap deliberately. This makes a rule file useful for repeated exports with consistent schemas, but it is not a guarantee that the new file has the same semantics or contains valid records.

Field names and constant values can be sensitive even without the table itself. Review a rule file before sending it to another person or storing it in a shared folder. A rule file also does not preserve the full multi-step memory session as a restorable project. Keep exported data separately when you need a durable result. Navigation within the site preserves the active session, while reloading or closing the page clears it in this version.