pandas Excel usecols: tracing renamed headers back to the selected sheet columns
Provide a mapping that preserves the user's selected coordinates and business meaning, rather than simply stripping a suffix from the column name.
pandas Excel usecols: tracing renamed headers back to the selected sheet columns
A suffix appearing after selecting only part of a sheet can reflect duplicate-name handling outside the selected range. Removing or replacing the display name alone can merge distinct source columns into one field. An address-based header mapping provides the basis for choosing the consumer's name safely. [1, 2, 3]
What the primary source establishes
A report shows read_excel adding a suffix even when the selected usecols range contains no duplicate names, because another column elsewhere in the sheet shares the name. In its C:D example, col 2 becomes col 2.1. [1]
The handover decision
Compare each selected address's original and parsed headers. Rename only through a mapping approved by the user. Keep sheet_column_address, source_header, parsed_header, and approved_header together so that similarly named columns are not silently treated as the same source. [1, 3]
A reference defined before the output
In the illustrative C:D selection contract, column D is named col 2. A consumer mapping that preserves col 2 satisfies the contract; col 2.1 does not yet match it. Do not remove suffixes globally before confirming the selected address, because another column may legitimately need a different identity. [1, 3]
The scoped Truthound check
Truthound Core 3.3.2 can compare an address-bound approved_header with the actual consumer header using ColumnPairValidator. Retain the sheet and column address instead of relying on names for uniqueness. Removing a suffix is an approved mapping transformation, not a validation result in itself. [1, 2, 3]
Limits and alternative choices
Different columns can share a name across the sheet, so name equality does not establish column identity. Identify columns by address and assign separate unique consumer-facing names. If the consumer contract already adopts pandas' suffix as an identifier, restoring the source name would itself be a contract change. [1, 3]
Application conditions
Prepare source, parsed, and approved headers for the same selected sheet-column address.
Alternatives
Identify columns by address and assign separate unique consumer-facing names.
Limitations
Different columns can share a name across the sheet, so name equality does not establish column identity.
Evidence
- Event checked
- Subject checked
- Modified
