JOIN USING

Non-Key Duplicate Names Stay Distinct

Duplicate non-key names are qualified.

highlighted = computed this step

Non-key duplicate names

If a non-key column name appears on both sides, the renderer keeps the side visible.

non key duplicates\text{non key duplicates}

Read duplicate names

There is 1 duplicate non-key name, so the output labels use left dot and right dot names.

duplicate names=1\text{duplicate names}=1

JOIN USING examples are tiny finite-table transforms; SQL dialect completeness, type coercion, optimizer behavior, execution cost, indexing, output ordering guarantees, and database-product claims are out of scope.

Non-key duplicate names stay distinct: left tableorder_idcustomer_idnote110new220rush320gift4NULLdraft Non-key duplicate names stay distinct: right tablecustomer_idnamenote10Adavip20Benstandard20Bennyduplicate30CyprospectNULLUnknownmissing candidate key comparisonsleftSourcerightSourceleftKeyrightKeymatched001010yes011020no021020no031030no0410NULLno102010no112020yes122020yes132030no1420NULLno202010no212020yes222020yes232030no2420NULLno30NULL10no31NULL20no32NULL20no33NULL30no34NULLNULLno JOIN USING output rowscustomer_idleft.order_idleft.noteright.nameright.note101newAdavip202rushBenstandard202rushBennyduplicate203giftBenstandard203giftBennyduplicate ON join contrast rowsleft.order_idleft.customer_idleft.noteright.customer_idright.nameright.note110new10Adavip220rush20Benstandard220rush20Bennyduplicate320gift20Benstandard320gift20Bennyduplicate JOIN USING factsfactvalueleftRowCount4rightRowCount5candidatePairCount20matchingPairCount5matchedLeftRows3duplicateMatchRows2nullKeyCandidatePairs8usingKeyAppearsOnceyesduplicateNonKeyNames[note]outputColumnCount5onOutputColumnCount6rowOrderleft_source_then_right_source

Qualified non-key labels

The two note columns stay separate as left dot note and right dot note.

qualified labels\text{qualified labels}

Summary

JOIN USING merges only the using key in this scoped model.

only key merged\text{only key merged}