Dremio keeps telling "Detected schema change"

1

Dremio version: 21.2.0-202205262146080444-038d6d
dremio.conf:

paths: {
  # the local path for dremio to store data.
  local: "/var/lib/dremio"

  # the distributed path Dremio data including job results, downloads, uploads, etc
  dist: "file://"${paths.local}"/pdfs"
}

Hi, when I format my data, Dremio preview the data correctly. But after format and query the data, I get no data. Checking the log and I see this

{"log":"2022-09-23 04:55:02,958 [e1 - 1cd2c3d8-99ad-e1d8-ab96-8d6076079a00:frag:3:0] INFO  c.d.s.e.f.FragmentStatusReporter - 1cd2c3d8-99ad-e1d8-ab96-8d6076079a00:3:0 scheduling weight 1: State to report: FINISHED\n","stream":"stdout","time":"2022-09-23T04:55:02.958714018Z"}
{"log":"2022-09-23 04:55:02,978 [e0 - 1cd2c3d8-99ad-e1d8-ab96-8d6076079a00:frag:2:0] INFO  c.d.s.e.f.FragmentStatusReporter - 1cd2c3d8-99ad-e1d8-ab96-8d6076079a00:2:0 scheduling weight 1: State to report: FINISHED\n","stream":"stdout","time":"2022-09-23T04:55:02.979106482Z"}
{"log":"2022-09-23 04:55:02,982 [e3 - 1cd2c3d8-99ad-e1d8-ab96-8d6076079a00:frag:1:2] INFO  c.d.s.e.f.FragmentStatusReporter - 1cd2c3d8-99ad-e1d8-ab96-8d6076079a00:1:2 scheduling weight 1: State to report: FINISHED\n","stream":"stdout","time":"2022-09-23T04:55:02.982935439Z"}
{"log":"2022-09-23 04:55:02,983 [cache-async-io-24] INFO  c.d.e.s.p.MutableParquetMetadata - Created parquet meatdata with row group size 1 for file /default/staging/pvoil/pvoil_price/2019.parquet.gzip\n","stream":"stdout","time":"2022-09-23T04:55:02.983856393Z"}
{"log":"2022-09-23 04:55:02,984 [e1 - 1cd2c3d8-99ad-e1d8-ab96-8d6076079a00:frag:1:0] INFO  c.d.e.s.p.UpPromotingParquetReader - F[setupMutator] Footer size is 1 for file /default/staging/pvoil/pvoil_price/2019.parquet.gzip, current row group index is 0\n","stream":"stdout","time":"2022-09-23T04:55:02.985064205Z"}
{"log":"2022-09-23 04:55:02,993 [e1 - 1cd2c3d8-99ad-e1d8-ab96-8d6076079a00:frag:1:0] INFO  c.d.e.s.parquet2.ParquetRowiseReader - Detected schema change. Not initializing further readers.\n","stream":"stdout","time":"2022-09-23T04:55:02.993332845Z"}
{"log":"2022-09-23 04:55:02,993 [e1 - 1cd2c3d8-99ad-e1d8-ab96-8d6076079a00:frag:1:0] INFO  c.d.e.s.p.MutableParquetMetadata - Removing row group index 0 for file /default/staging/pvoil/pvoil_price/2019.parquet.gzip\n","stream":"stdout","time":"2022-09-23T04:55:02.993609378Z"}
{"log":"2022-09-23 04:55:02,995 [e1 - 1cd2c3d8-99ad-e1d8-ab96-8d6076079a00:frag:1:0] INFO  c.d.s.e.f.FragmentStatusReporter - 1cd2c3d8-99ad-e1d8-ab96-8d6076079a00:1:0 scheduling weight 1: State to report: FINISHED\n","stream":"stdout","time":"2022-09-23T04:55:02.995589707Z"}
{"log":"2022-09-23 04:55:02,993 [e2 - 1cd2c3d8-99ad-e1d8-ab96-8d6076079a00:frag:1:1] INFO  c.d.s.e.f.FragmentStatusReporter - 1cd2c3d8-99ad-e1d8-ab96-8d6076079a00:1:1 scheduling weight 1: State to report: FINISHED\n","stream":"stdout","time":"2022-09-23T04:55:02.995806094Z"}
{"log":"2022-09-23 04:55:03,004 [e0 - 1cd2c3d8-99ad-e1d8-ab96-8d6076079a00:frag:0:0] INFO  c.d.s.e.f.FragmentStatusReporter - 1cd2c3d8-99ad-e1d8-ab96-8d6076079a00:0:0 scheduling weight 1: State to report: FINISHED\n","stream":"stdout","time":"2022-09-23T04:55:03.005114784Z"}

Image: Dremio previews data correctly

image
image1197×406 23 KB

Image: After that, Dremio shows no data

image
image1814×715 35.2 KB

2

Hi @chulucninh09,

could you please check that all your files share a same schema - not only json, but also regarding json structure ?
Dremio use only a subset of the data to infer schema, and then can be stuck when parsing all the data corpus if it finds out that a file does not share the same schema.

Hoping it helps,

3

@fetanchaud @chulucninh09 Schema mismatch should not cause an empty result, are you able to send me the profile and full log file?