A few comments on this family and research in general
If your records are pre1900 accept that there will be spelling variations e.g. Neal & Neill; most people were illiterate and someone else wrote their details on forms, this leads to many variations of the same names based on accents, locations, etc.
The same can be said of ages, years of marriage, etc. these can be out by several years when people had nothing to officially use. When searching the records you can find a range of 10 years for dates of birth, marriage, and so on; most interesting is when you find dates of baptism before the official date of birth!! This can arise where the family was avoiding a fine for a late notification and they gave a date within three months of the date they were regisitering the birth, whereas the church would have recorded the actual date of baptism in their books!! Makes for difficult researching sometimes.
Seeing as Family Search and Roots Ireland are giving the same year for the marriage of Peter Sweeney & Mary Neil I would accept there is no error in the transcript; there may be for Mary's birth or indeed it may be correct!
If you could access the original church marriage record there may be a note referring to the status of the marriage (e.g. dispensation or otherwise) or the original baptism record to see if the year was correct.
Many people were listed on census returns as visitors when they were in fact relatives; I have found that in-laws and step-children in particular were not seen as "related to the head"; If your Mary Sweeney listed with the Rourkes is the correct age, shown as married, etc and in other ways matches I would be happy to keep her a very strong possiblity
I hope this helps you think it through a little more