ok,thanks.
My other question is can Plan be reused? I'm going to use Acero to read Parquet files stored using Icerberg. I will have two tables, a big table and a small table. The small table and the large table have one column with the same UUID and the other columns are different. The parquet file for A small table is A list named A, and the Parquet file for A large table is A list named B. The relationship between these two lists is that each file in list A has A UUID that is A subset of the UUID of several files in B. For example, A file in A has A UUID of 200-1000, while in B there may be three files with UUID of 0-400,400-800, and 800-1200. At this time, I will first read out the file in A, and then read out A file in B, and then use Acero to do join and filter operations on them. This is repeated three times, and only the sourceNode changes during the three sessions. Is there a way to reuse the rest of the nodes? 1057445597 [email protected] ------------------ ???????? ------------------ ??????: "user" <[email protected]>; ????????: 2022??8??23??(??????) ????2:24 ??????: "user"<[email protected]>; ????: Re: Ask some questions about Acero Well, all code has potential bugs :). We generally do not lower our quality standards for experimental code. So it should still be unit tested, reviewed, etc. The term is primarily meant to indicate that the interface could be modified later. Acero, in particular, has quite a few tests in place. There are some known bugs, the most critical of which is probably a potential crash-on-shutdown bug, which is being worked on. On Mon, Aug 22, 2022 at 7:56 AM Aldrin <[email protected]> wrote: I think it means both Aldrin Montana Computer Science PhD Student UC Santa Cruz On Mon, Aug 22, 2022 at 2:53 AM 1057445597 <[email protected]> wrote: I noticed that ??Acero is experimental and a stable API is not yet guaranteed?? Does this mean that Acero's interface could be modified later, or are there potential bugs in the functionality? 1057445597 [email protected]
