Processing large results

It’s possible to iterate over large result without loading it as a whole into a memory. The ResultStream<'t> class allows to read data sequentially. It keeps open reader until the result is read to the end, or the stream is explicitly disposed. To use the streaming functionality, it’s enough to define query function with ResultStream as a result, e.g:

let getAllComments: int -> AsyncDb<ResultStream<Comment>> = 
    sql "select c.id, c.postId, c.parentId, c.content, c.author, c.createdAt 
         from comment c join post p on c.postId = p.id 
         where p.blogId = @id"

Since it implements seq interface, it can be iterated as any other collection:

asyncdb {
    for c in getAllComments blogId do
        ...
} |> runAsync 

When iterating whole result set, simple function call is enough, but if we intend to break iteration, the preferred way is to enclose the result stream in a use statement:

asyncdb {
    use! comments = getAllComments blogId 
    for c in comments do
        ...
} |> runAsync