Mapping recipes

The config matrix lists every MapEngine knob and vocabulary, and Mapping explains how the scorers combine. This page is the task-shaped shortcut: a few complete snippets for the jobs you actually reach for.

Map a messy export to your canonical schema

You have a source file whose column names don’t match your target. You want the best 1:1 mapping and the renamed frame.

import infermap
import polars as pl

result = infermap.map("crm_export.csv", "canonical_customers.csv")
for m in result.mappings:
    print(f"{m.source} -> {m.target}  ({m.confidence:.0%})  {m.reason}")

renamed = result.apply(pl.read_csv("crm_export.csv"))   # rename source columns to target names

Each mapping carries a confidence in 0.0..1.0 and a human-readable reason, so a low-confidence pair is easy to spot and override before you apply it.

Reuse a saved mapping as a validation gate

You have recurring files from the same source. You want to pin the mapping once and fail fast when a new drop is missing a required field.

result = infermap.map("crm_export.csv", "canonical_customers.csv")
result.to_config("crm_mapping.yaml")            # commit this

Then validate each incoming file against the pinned config — non-zero exit on a missing required field makes it a CI gate:

infermap validate new_drop.csv --config crm_mapping.yaml \
  --required first_name,last_name,email --strict

Apply the pinned mapping without re-inferring:

infermap apply new_drop.csv --config crm_mapping.yaml --output clean.csv

Add a custom alias or domain pack

You have source names your target schema doesn’t know about (cust_ref really is customer_id). You want InferMap to resolve them without lowering min_confidence.

# mapping_rules.yaml — engine config (domains, aliases, scorer weights)
domains: [finance]
aliases:
  customer_id: [cust_ref, acct_no]
  email: [email_addr, contact_email]
scorers:
  FuzzyName:
    weight: 0.4

from infermap import MapEngine

engine = MapEngine(config_path="mapping_rules.yaml")   # loads domains + aliases + scorer weights
result = engine.map("crm_export.csv", "canonical_customers.csv")

Aliases feed the AliasScorer (weight 0.95), so a known alias resolves with high confidence instead of relying on the fuzzy-name tie-breaker. (Note this is the engine config file — distinct from the saved mapping an earlier result.to_config(...) writes, which infermap apply/from_config consume.)

Get started

Concepts

GoldenMatch

GoldenCheck

GoldenFlow

GoldenPipe

GoldenAnalysis

InferMap

SQL extensions

Reference

Research

Map a messy export to your canonical schema

Reuse a saved mapping as a validation gate

Add a custom alias or domain pack

​Map a messy export to your canonical schema

​Reuse a saved mapping as a validation gate

​Add a custom alias or domain pack

Map a messy export to your canonical schema

Reuse a saved mapping as a validation gate

Add a custom alias or domain pack