A lightweight Stata command that augments merge by checking what happens if you merge on unique combinations of the key variables.
net install smartmerge, from("https://raw.githubusercontent.com/rdahis/smartmerge/main/smartmerge") forceWhen either the master or using data contain duplicate keys, merge behaves as documented but the resulting observation counts can be hard to interpret. smartmerge shows you – in a single step – what the merge result would be if duplicates were dropped first so you can see whether duplicates are the root cause of weird merge counts.
smartmerge mirrors the syntax of Stata's native merge. Everything after the command name is passed through verbatim to merge.
smartmerge 1:m id using survey2, keep(match master)The final dataset in memory is identical to what you would get from the call to merge alone. A diagnostic table for the duplicate-free merge is printed to the Results window.
Feel free to open issues or pull requests. This repo is intentionally minimal – PRs that add tests, real-world examples, or performance improvements are very welcome.