When not using the rs.each block to read each row sequentially, it is seemingly necessary for rs.move_next to be called before the returned data can be accessed.
This doesn't seem to be documented in any visible fashion and lead to half an hour of staring at my query trying to figure out what was wrong with my code.
As an example, the following causes read the column <column> returned a Nil but a Int64 was expected.:
DB.open <uri> do |db|
db.query "SELECT <column> FROM <table> WHERE <column> = <value>" do |rs|
puts rs.read(Int32)
end
end
By contrast, this code returns the expected value, even if the table only has a single row:
DB.open <uri> do |db|
db.query "SELECT <column> FROM <table> WHERE <column> = <value>" do |rs|
rs.move_next # <-- Added line
puts rs.read(Int32)
end
end
As I mentioned before, this seems to imply that calling rs.move_next or use of the rs.each block is mandatory for the query results to be populated, which unless I'm misunderstanding something seems a unintuitive.
If this intended, please document it somewhere easily visible to save future trial and error sessions from other potential users.
When not using the
rs.eachblock to read each row sequentially, it is seemingly necessary forrs.move_nextto be called before the returned data can be accessed.This doesn't seem to be documented in any visible fashion and lead to half an hour of staring at my query trying to figure out what was wrong with my code.
As an example, the following causes
read the column <column> returned a Nil but a Int64 was expected.:By contrast, this code returns the expected value, even if the table only has a single row:
As I mentioned before, this seems to imply that calling
rs.move_nextor use of thers.eachblock is mandatory for the query results to be populated, which unless I'm misunderstanding something seems a unintuitive.If this intended, please document it somewhere easily visible to save future trial and error sessions from other potential users.