Searching for place names containing numerals produces unexpected results.
Often there are place names that contain numerals.
- Justice Precinct 1, Nueces, Texas, United States
- Justice Precinct 2, Lubbock, Texas, United States
- Justice Precinct 3, Fannin, Texas, United States
- Justice Precinct 4, Runnels, Texas, United States
- Commissioner's Precinct 4, Jackson, Texas, United States
- Tract 7, Houston, Justice Precinct 1, Harris, Texas, United States
- Election Precinct 18 South of Rocky Ford, Otero, Colorado, United States
- Election Precinct 21, Jefferson, Alabama, United States
Using the 'Residence' life event, entering an exact search for the following place name criteria will work as expected: "*precinct ?,*" or "*precinct 4,*" or " 4,*"
An exact search for "* 1124,*" will also produce valid results.
However, an exact search for "* 1124*" performs in an unexpected manner, and produces 604,737 results, which for reasons unknown to the unwitting researcher, are mostly from the 1940 US Census.
In another question: https://community.familysearch.org/en/discussion/106058/why-are-certain-place-names-prefixed-with-a-number I cited a result containing "1142 Shrewsbury, Salop, Eng"
So there should be a valid search criterion to allow the researcher to locate a record containing that exact content. It seems that there is not.
Similarly, a previouly seen result which contains "225, Saint Michael, Gloucester" should be possible to locate using a place name search. Again, It is not. Performing a search using the criterion "*225, Saint*" produces 4 results all from Paris, France containing the address "Fg St Martin 225". Removing the asterisk from the start of the criterion produces no results at all.
Using the exact criterion "*225*" produces 8,510,473 results from the 1940 US Census.
In the results from the 1940 US Census: None of the viewable information in any of the records in the results contains the search criterion. Yet the majority of results are from this collection. It seems that either the query processor is modifying the criterion, or the search engine is locating data that is not viewable, such as an index number.
Either way, the need to search for numerals in place names is real, and the query processor and search engine should produce understandable results. If the user recieves a huge tranche of results in response to a query, they might begin to wade through all of them forlornly believing that their objective is in the list when in fact it is not.
Answers
-
Yes, some explanation on wildcards/acceptable regex constructions is needed - otherwise these 'features' remain 'hidden' to the 'regular' searcher.
I guess keep eyes peeled for Rootstech - otherwise I don't know why this 'new feature/what's changed' education doesn't occur prior, during, and after production releases.
0 -
After some experimentation adjusting the exact search criterion to reduce the number of results, I was able to determine that the query with the form: "*231124*" will locate address content. In this case the search engine has determined that the exact criterion is contained within the address: "2311 24th St." It is not. This is not correct behaviour.
Also, the exact queries: "*d1124*" and "*e1124*" will locate the following:
Residence Place: 16 Ward, 1124 or Address: Rue de Grenelle 1124 or Address: 1139 Conn. Ave., 1124 11 St. S.W.
These are not exact matches.
I have yet to determine how the search engine is matching "*??????????????????????????????????????1124*" and still getting 41 results, but is seems to be clear that delimiters and punctation marks are ignored, so any combination of separate words or numbers including spaces or punctation marks that contains 1 1 2 and 4 in that order will match.
1