Two trees that overlap, and you want one. A cousin's file and yours. An old file and the export from a website. Two branches researched separately. Merging is where most trees pick up their duplicates and their disconnected islands, so it is worth doing slowly.

When not to merge

If the second tree is mostly unsourced and yours is not, merging pulls unproven people into your work. Keep it as a reference file instead; open it in GED-VIEW when you need to look something up.

If the two trees disagree on facts you care about, resolve those first. A merge does not settle arguments; it hides them.

If you just want to see the two side by side, open each in its own tab.

The workflow

  1. Open the main file in GED-VIEW; it is file A. Open Tools > Compare files and pick the second file as file B to see what is in one and not the other, and where the same people appear in both. That shows how big the overlap is and where the joins will be. The compare guide explains what it reports.
  2. Open Tools > Merge trees. File A is the open file; file B is read privately and is never saved or opened. The tool proposes who looks like the same person: only pairs whose surname, given name, birth and death years and sex agree wherever both files have them (a shared unique ID or reference number counts on its own), with the reasons listed on each card, and which sources, notes, media and repositories match by content.
  3. Decide every pair yourself. Nothing is pre-selected: each pair is "The same person", "Different people" or "Decide later". One button marks the very sure pairs as the same at once, and every card still shows its answer. Pairs the tool missed can be added by hand.
  4. Decide what goes into each merged record. Every piece of information (each name, event, note, citation, media link, identifier) is one row: identical on both sides, only in one file, or different. A difference is your call: keep file A, keep file B, keep both where GEDCOM allows several, or neither. Records only in file B come in unless you leave them out.
  5. Check the families the tool merges (the same couple in both files) and the issues it found. A decision that would make someone their own ancestor blocks the result until it is changed.
  6. Read the Review, then choose an ending: Download the merged file, Open it in the app as a new file next to the others, or Preview it: every page colors what comes from file B in green and what file A's records took in in amber, and Accept in the file menu saves it as a new file.

Neither input file is modified. A long review is saved in the browser on every click, so it can be picked up after a reload.

Things that always need a look

  • Same-name children in one family. The tool never proposes twins or siblings as the same person, but check the family cards when two Johns appear; if the first died young they are two people.
  • Women under maiden and married names. Different names, same person. The tool never proposes them (the surnames differ) unless both records carry the same unique ID, so add the pair by hand.
  • Places. Two files spell places differently, so after a merge every variant of every town is present. The place standardization guide covers cleaning that up with the Standardize places and dates tool.
  • Sources. Two files that cite the same census will have two source records for it unless the records match by content. Find duplicates can merge any that turn out identical afterwards; differently worded ones need a decision by hand.

Next

Compare first, and see how much would actually be joined.