From WeRelate
WeRelate (through volunteers and people working on their own information) seeks to eliminate duplicate Person and Family pages, resulting in one page for each unique person and one page for each unique family. This is a key goal of WeRelate.
This help page is primarily for merging duplicate Family or Person pages. Nonetheless, it is also possible to merge duplicate Place or Source pages; see the FAQ below for information on merging places or sources.
Use the Talk page for asking questions not covered here or for additional support.
Merging duplicate Family or Person pages
Merging duplicate Family or Person pages involves three things:
- Finding pages that are likely to be duplicates.
- Comparing the pages that are likely to be duplicates side-by-side and deciding which pages to merge.
- Deciding which pieces of information (names, events, sources, images, notes, and text) to keep on the merged page.
Finding pages that are likely to be duplicates
The first thing to do is to find pages that are likely to be duplicates. You can do this in one of two ways:
- click on Show duplicates in the MyRelate menu to see a list of families in your tree that are likely to have duplicates, and click on one of the families in the list, or
- navigate to a Person or Family page that you think might have duplicates and select Find duplicates in the More menu at the top of the page.
Each of these takes you to a Search results page, where you see a list of other pages that might be duplicates. Check the boxes next to the pages you want to compare and press the Compare button. This takes you to the Compare pages to merge screen, where you see the pages you've selected side by side.
Another approach is to navigate to a Person page that lists multiple parents or spouses and select Compare parents or Compare spouses in the More menu, or navigate to a Family page that has multiple husbands or wives and select Compare husbands or Compare wives in the More menu. This also takes you to the Compare pages to merge screen.
Comparing pages to merge
In the Compare pages to merge screen, you compare vital information from each of the pages to merge in side-by-side columns.
- Green boxes mean the information is specific and matches exactly.
- Yellow boxes mean the information is missing some pieces (e.g., a date that is a year only or a place that is just a country or a US state) or the information is a partial match (e.g., names that sound similar but are spelled differently, or places where one is a county and the other is a town in the county).
- Red boxes mean the information differs.
Merging guidelines
Wikis are about collaboration and working together, sharing what we know. Having duplicate pages makes this less effective. That said, we don't want to offend people who have contributed extensive research to our knowlegde base. Listed below are some basic rules of thumb for merging pages when you are not the author of both.
- WHEN IN DOUBT DO NOT MERGE. You can send the persons watching the pages an email or leave a message on the talk page and ask them to work it out.
- MERGING PAGES WITH MINIMAL COMMON INFORMATION IS NOT A SERVICE TO ANYONE.
- Remember you are smarter than the computer. And, just because two pages meet the criteria to be on the duplicate list, doesn't mean that they are duplicates. There are just some things we can't teach computers to do for us. That's why we have people make the final merge decisions.
- If you are merging a person page with significant data or documentation with a page that has only names that match and one event with same date and place without any other disagreement--merge.
- If you are merging a family page with significant data or documentation with a page that similar spouse names in agreement (one spouse's name may be unknown) and a match on the marriage date and place or a match on a child--merge.
- If you are merging two pages with minimal information it may be best not to merge. There are five Lewis Green Caddells all contemporaries and all buried in the same cemetery. Two were married to a Mary Elizabeth. The problem here is that there may be several couples with similar names, living in the same area, and each couple may have had at least one child with the same name, such as Daniel Caddell (1) and Daniel Caddell (2) without Dan's birth date and place in agreement, there is no way to know we have the same Lewis Caddell and Mary Elizabeth Unknown. Sometimes people post pages with minimal information as place holders, hoping that they or someone will eventually solve the puzzle and fill in the page. It doesn't help to merge until we have more information.
- If you are merging a page with significant data or documentation with a page that has minimal info and that info is only a partial match--send the persons watching the documented page a message and come back to it in a week or so. If they haven't come forward, responded or otherwise taken an interest, merge the pages if you think it is appropriate. But give people who seem to have real interest in their data the chance to make the merge decision. They may know more information about the people than has been posted. Some users purposely use multiple spelling of the surname for members of the same line as a way of recording the alternate spellings. Those users are the only ones that know would consider Smith and Schmidt a true duplicate.
- If you are merging 2 pages with significant data, the point system is best. When deciding whether to merge two individuals, a general rule of thumb is to add two points for each green item, one point for each yellow item, and subtract two points for each red item. That is, give 2 points for each data element that is an exact match (green on the match screen), such as given name, surname, birth date, birth place, death date, death place, child's name, parents' name. If it is a vague but similar reference such as a US state or a country, a year or year range instead of an actual date (yellow on the match screen), give it only one point. If you have a disagreement (red on the match screen), subtract 2 points. Keep in mind that humans are better at spotting matches than the computer.
- Generally, if 8 points match, you can be fairly certain it is the same person or family. With the more ancient files, six might be sufficient. This rule isn't always correct, but it's a good starting point. Again, when in doubt don't merge, email the persons watching the page or leave a message on the talk page and let them work it out. Come back in a week and merge if that seems appropriate.
How to merge
- Choose the pages to merge. Check the boxes above the pages you want to merge, then click on the "Prepare to merge" button at the bottom of the page.
- Choose the information to keep. The merge page will show you all of the names, events, sources, images, notes, and text on each of the pages you've chosen to merge, and you will need to decide which pieces of information to keep.
- The merge target is the page that the other pages will be merged into. It defaults to the earliest-created page. Unless this page has the wrong name in the title, you should keep this page as the merge target.
- To choose the information to keep on the merge target, check the boxes to the left of each item to keep. Items from the merge target that are kept will appear on the page before items imported from other pages; text from the merge target will appear above text imported from other pages.
- If you keep multiple birth, christening, death, or burial events, the other events will be shown as "alternate" events.
- If you have the same date and place as an existing event but your event has a (non-gedcom) source that you want to attach to the existing event, leave your event unchecked and check the box next to your source. Your source will be attached to the existing event. or
- You don't need to worry about source & note ordering; sources & notes will be re-ordered automatically.
- Because the computer is a computer, it will often mark items as "different" and therefore keep them by default when they are in fact the same as other information. For example, a county name might be misspelled, or one user might have entered 1743 - 31 - 5 instead of 31 May 1743. Best practice is to keep only the best piece of this mostly-matched data. If the source on the other data is better, keep that (it will be added to the matching event upon the merge). If the data is different, generally keep the different information unless it is obviously wrong (i.e. born 1600 in Massachusetts) and follow wiki etiquette for dealing with questionable information.
- Do not keep user ID numbers, reference numbers, and other single-user identifiers of the information. (But do keep Ancestral File numbers or other universal ID numbers, as they can be used to identify duplicates or find that individual in the AF.)
- Do not keep MySources that reference gedcom, .ftw, etc. files.
- Once you've chosen which pieces of information to keep, press the Merge button at the bottom of the page to merge the pages. After a short pause you'll see a list of the pages that were merged. If you're merging well-populated pages with lots of sources, images, notes, and/or text, it's a good idea to view the merge target and edit it in case it needs some tweaking.
- When merging families, children not selected for merging will be added to the merged family along with the merged children.
Handling Ambiguous Spouses Resulting from a Merge
When merging one or more seemingly duplicate pages within a particular ancestral line, you may often encounter multiple candidates for the husband or wife of a family (typically the wife, due to surname issues). How do you choose? This section seeks to guide you through this process.
How does it happen?
The bulk of duplicates on WeRelate occur from uploaded GEDCOMs when the data within that GEDCOM contains either:
- duplicates within itself (i.e., a single GEDCOM contains multiple instances of a given person). This can happen:
- If a child is incorrectly assigned to two different families, generally with the same husband.
- If a researcher is too quick to accept a possible spouse suggested by a search tool
- If a researcher is too quick to accept the work of others, who made either of the above mistakes.
- a duplicate of a Person or Family already on WeRelate at the time of the GEDCOM upload.
(Note: such duplication rarely happens when Person or Family pages are created manually by hand.)
How to correct the duplication?
When multiple spouses are paired with the same individual, first determine:
- Is this one marriage or multiple marriages?
- Are they multiple candidates for the same marriage, or are they both spouses of the same person via multiple marriages (or even a combination of both, e.g., one spouse for one marriage, two other spouses for another marriage)?
Just doing a straightforward merge may eradicate the subtle differences between these different cases.
The person doing the merging (the merger) must try to determine, as much as possible, how many distinct marriages are represented, so that they can end up with one Family page for each actual marriage to allow children to be placed in the correct marriage regardless of who the actual spouse is later identified as. Two candidates for being a spouse in a marriage, must remain as separate Person pages to allow their potentially different parents and vital data to be collected separately from other candidates.
Ideally, you can inspect the contents and see sources that explain the correct relationships. Less ideal (but not unreasonable when source information is missing or unclear) compare marriage dates of spouses with birth dates of children to see if any children were incorrectly attributed to both families (in that case, just make the change and add a note that this is probably the actual reality).
Additional options could include:
- Just allow the different names to continue to appear as alternative spouses. This can be a reasonable choice when it's not yet clear which children belong to which spouse (usually but not always the mother). The family will continue to appear in "duplicate" searches, and others researching the said lines find this, they may add their sourced information and resolve the situation more quickly.
- Add a "Disputed Lineage" section to the pertinent Person or Family page(s), summarizing what's known about the (duplicate) candidates.
- When it's clear that there were two distinct families (e.g., a husband with more the one spouse), be sure to have two separate Family pages. For any children that are known (and documented) to be associated with a specific set of spouses, assign them appropriately.
- When the precise parentage of children is undetermined, make a good educated guess but be sure to include a "disputed lineage" section on the Person pages of said children.
- A "disputed lineages" section on the pertinent Family page would also describe alternative possibilities for the actual family.
- In some situations, where you have mangled names or plainly scrambled family relationships, and no source citations whatsoever, it is not unreasonable to nominate such pages for speedy delete. After all, pages can be easily recreated when good information is available.
Things to Consider
As you evaluate which of the different approaches to pursue, consider
- What is the most explicit/obvious way to present the uncertainty?
- Might the result incorrectly be interpreted as yet another duplicate by the automatic duplication detection mechanism of WeRelate? If so, how can you edit the pertinent page(s) to prevent this?
- What would a GEDCOM upload of the affected person/family pages look like?
- Would the ambiguous situation be apt to be retained or lost in transition from WR to another application? (The other application might be WR, as a result of an upload of off-line work).
When the results include conflicting data
Unless you have personal knowledge, which you should add sources for, leave conflicting data. Consider placing using the "Questionable data" template that looks like this:
Questionable Information Found
A sentence describing the conflict. Documentation needed.
Leave all possibilities for researchers familiar with the person or family to identify what is wrong and remove it. The goal is to get all information about one person, or one family, on a single page, even if that page has multiple versions of various facts.
Frequently asked questions
Why are some pages semi-protected?
Semi-protected pages cannot be changed during a merge or a GEDCOM update. To change them, you need to navigate to the page and edit in in the normal way.
There are two reasons a page is semi-protected:
- Pages with more than five people watching them are semi-protected under the (possibly false) assumption that information worth adding to the page has been added already.
- Pages for famous people with a wikipedia link are semi-protected under the (possibly false) assumption that information for famous people is more likely to be correct on Wikipedia than elsewhere.
If you have additional or corrected information to add to a semi-protected page, please add it by editing the page.
What happens after a merge?
- Everyone watching the duplicate page will be added to the list of people watching the merge target.
- Everyone watching either page will get an email notifying them that the page has been changed.
- Everyone's network will be updated to include the users that they are now related to.
- All the links to the duplicate page on other Person/Family pages will point to the merge target.
- Any other links, such as those manually entered in notes, will continue to work and will be redirected to the merge target.
- The duplicate pages will no longer appear in search results.
What if I don't agree with a merge?
- If people were merged incorrectly--that is, two different people were merged into one person, you can undo the merge. See below for instructions.
- If two pages that are actually the same person were merged, then they should stay merged. Any corrections should be made on the newly merged version of the page. If there are disagreements over information that should or should not be included, use the Talk page to discuss what should be properly included.
Instructions for Unmerging pages
If pages are merged incorrectly and you want to undo the merge, here's what to do:
For each page involved in the merge, both Family pages and People pages, both the merge targets and the duplicates:
- Navigate to the page. For redirected pages, make sure that you're looking at the redirected page by clicking on the "redirected from page title" link at the top of the merged page. You can see a list of the pages that were involved in the merge by viewing your "Contributions" screen (click on MyRelate, then Contributions).
- Click on the "History" link.
- Click on the time-and-date link on the revision just before the merge occurred (the line just below the line with the merge comment).
- Click on the "Edit" link.
- You should see a warning about editing an out-of-date revision of the page. Ignore the warning and save the page. Add a summary comment about unmerging.
If you want to check to make sure that you've unmerged everything correctly, after you've edited all of the pages navigate to each page again:
- Click on the "History link.
- Click on the radio button (the circle) next to the revision just before the merge occurred and then click on the "Compare selected versions" button.
- The only changes you should see between the pre-merge revision and the current revision should involve re-ordering data elements and standardizing place links. Person and family links should be the same between the two revisions.
If you need help, please leave a message on User talk:Dallan.
Note: If a page that was redirected has a talk page associated with it, you'll see a "View the talk page" link instead of the "redirected from page title" link at the top of the merged page. This is annoying, and it's something I'm going to fix in a few weeks. In the meantime, if this happens to you, add "?redirect=no" (without the quotes) to the end of the URL line and press enter to navigate to the redirect page.
What to do when similar pages should not be merged
If you come across two Family pages that appear similar but should not be merged, and you want to let others (and the site) know that they should not be merged, edit the Talk page of one of the families and add a template
- {{nomerge|Family:Title of the other family}}
somewhere on the page. You can do this for Person pages as well.
Pages marked this way will not appear in the duplicates list next day.
How to merge duplicate Place pages or duplicate Source pages
- Decide which page to keep. We'll call that page the target and the other page the duplicate.
- Edit the target page. Copy over information from the duplicate that you want to keep into the merge target page.
- Edit the duplicate page. Erase everything in the big text box and replace it with
- #REDIRECT [[Place:Title of place target page]] or
- #REDIRECT [[Source:Title of source target page]]
That's it. The duplicate page will now automatically redirect people to the merge target page.
- NOTE: when pasting text into the brackets, be sure not to include any HTML code; for example if the target has a comma in the title (e.g., "Source:Ancient Historical Records of Norwalk, Conn."), make sure to type in a comma. Some paste action will paste HTML code and the redirect won't work; so if you see a redirect that looks like this after saving the edit:
- 1. #redirect Source:Ancient Historical Records of Norwalk, Conn.
- chances are there is some HTML code in the text on the edit page. Ideally, you want the result to look like this on the duplicate page:
MySource:Jillaine/Ancient Historical Records of Norwalk, CT<b>
Redirect page
[graphic of arrow] Source:Ancient Historical Records of Norwalk, Conn.
