You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The type change in #25 caused a lot of downstream type errors.
Previously in v1.1.1 there was only any[][] types returned to onComplete.
Then in v1.2.0 we updated the type to any[][] | Record<string, any> but any project which was passing an onComplete function with expected type any[][] was throwing a typescript error, and was not easy to fix.
In v1.2.1 I reverted the type change, but left the behavior of rowFormat: 'object' where it will return Record<string, any>[] instead of any[][]. This means the types are not consistent with behavior in that case.
I don't know how I want to fix this yet. But I refuse to break downstream projects that depend on the types.
I think I will start looking at making a new entry point to hyparquet. I want the rows to be returned as a promise instead of a callback. But this may also be a good opportunity to fix the awkwardness of the current parquetRead function.
The text was updated successfully, but these errors were encountered:
The type change in #25 caused a lot of downstream type errors.
Previously in v1.1.1 there was only
any[][]
types returned toonComplete
.Then in v1.2.0 we updated the type to
any[][] | Record<string, any>
but any project which was passing anonComplete
function with expected typeany[][]
was throwing a typescript error, and was not easy to fix.In v1.2.1 I reverted the type change, but left the behavior of
rowFormat: 'object'
where it will return Record<string, any>[] instead of any[][]. This means the types are not consistent with behavior in that case.I don't know how I want to fix this yet. But I refuse to break downstream projects that depend on the types.
I think I will start looking at making a new entry point to hyparquet. I want the rows to be returned as a promise instead of a callback. But this may also be a good opportunity to fix the awkwardness of the current
parquetRead
function.The text was updated successfully, but these errors were encountered: