20th: Some people make genealogy hard

I do so hate it when I cannot find someone who has to exist.
Last sightings were a marriage in Ontario, Canada in Peterborough 1893, followed by the births of two daughters, Edith and Gertrude, and death of wife Sarah (1899).
His name was supposedly Harry FAIRBAIRN, and he (they) should have appeared in either the Ontario deaths, or 1901 and subsequent 1911 census records.
No joy. Casting the net wider to outside Ontario, and even into the States still didn't come up with any likely candidates.
So I worked backwards to figure out which FAIRBAIRN family he belonged too, and quickly found that he was more usually known to his family as James Henry, and enumerated variously as Jas H, or James H with the surname as FAIRBAIRN or FAIRBURN. This latter is easily overcome by doing exact searches using wildcards, eg FAIRB*, but the forename variants are a little harder to cope with.
For the 1901 Ontario census I eventually resorted to simply leaving the name blank but putting in a birth year with +/- 5 years, and the township where I thought he would be, and scanning the entire 400 names for anything that could conceivably be a mis-indexed FAIRBAIRN. Still no joy.
By now I'd found that he had remarried (1899), as James Henry, to a widow Amelia COUTTS nee HUNTER, so it really seemed unlikely that he would be missing from both the 1901 and the Ontario deaths. But nothing anywhere in Canada to an Amelia with husband James or Jas or Henry or Harry, nor in the States.
I don't give up easily, so decided to try the known children. Gertrude had died at age 1, so it had to be Edith.
She did eventually provide the missing link I required. At age 17, in 1911, she had married in Renfrew Co, Ontario, "with parental consent", and gave her place of residence as Port Huron, Michigan, and her parents as Harry FAIRBAIRN and (incorrectly) Amelia HUNTER.
Working from the known to the unknown, always a good maxim, 1910 Michigan census finally provided what I'd been looking for, but even then only by putting in Edith with father Harry born Canada in the right timeframe, no surname. They were indexed as RAIBOURN. By 1920 Harry is indexed as "H FOUNTAIN", but as they still had son Percival aka Percy who was with them in 1910, with them in 1920, they were more easily found.

Even son Percival/Percy, supposedly born Ontario, I had missed in the birth registers first time through. His mother had been indexed as Ammelin, and father recorded as Henry J. And was he Percival? No, John P.
Those we chase around the records don't make it easy for us, nor do the transcribers.
How many forename variations? Jas, James, Henry, Harry, H, Henry J Coupled with FAIRBAIRN, FAIRBOURN, RAIBOURN, FOUNTAIN, FAIRBURN as indexed surnames, which makes strking the right combination of both names a bit of a challenge.
And after all that I probably still wont find a dna candidate for the line of Andrew FAIRBAIRN and Elizabeth HAGERMAN in the FAIRBAIRN Surname DNA project, but it sure wont be from lack of trying!