ARROW-5950: [Rust][DataFusion] Add logger

[Dandandan]

It seems useful to me to do some extra (debug, info, etc) logging.

I added to the benchmark.
Executing with RUST_LOG="debug" gives (next to sqlparser logs):

[2020-12-21T23:04:59Z DEBUG datafusion::execution::context] Logical plan:
     Sort: #l_returnflag ASC NULLS FIRST, #l_linestatus ASC NULLS FIRST
      Projection: #l_returnflag, #l_linestatus, #SUM(l_quantity) AS sum_qty, #SUM(l_extendedprice) AS sum_base_price, #SUM(l_extendedprice Multiply Int64(1) Minus l_discount) AS sum_disc_price, #SUM(l_extendedprice Multiply Int64(1) Minus l_discount Multiply Int64(1) Plus l_tax) AS sum_charge, #AVG(l_quantity) AS avg_qty, #AVG(l_extendedprice) AS avg_price, #AVG(l_discount) AS avg_disc, #COUNT(UInt8(1)) AS count_order
        Aggregate: groupBy=[[#l_returnflag, #l_linestatus]], aggr=[[SUM(#l_quantity), SUM(#l_extendedprice), SUM(#l_extendedprice Multiply Int64(1) Minus #l_discount), SUM(#l_extendedprice Multiply Int64(1) Minus #l_discount Multiply Int64(1) Plus #l_tax), AVG(#l_quantity), AVG(#l_extendedprice), AVG(#l_discount), COUNT(UInt8(1))]]
          Filter: #l_shipdate LtEq CAST(Utf8("1998-09-02") AS Date32(Day))
            TableScan: lineitem projection=None
[2020-12-21T23:04:59Z DEBUG datafusion::execution::context] Optimized logical plan:
     Sort: #l_returnflag ASC NULLS FIRST, #l_linestatus ASC NULLS FIRST
      Projection: #l_returnflag, #l_linestatus, #SUM(l_quantity) AS sum_qty, #SUM(l_extendedprice) AS sum_base_price, #SUM(l_extendedprice Multiply Int64(1) Minus l_discount) AS sum_disc_price, #SUM(l_extendedprice Multiply Int64(1) Minus l_discount Multiply Int64(1) Plus l_tax) AS sum_charge, #AVG(l_quantity) AS avg_qty, #AVG(l_extendedprice) AS avg_price, #AVG(l_discount) AS avg_disc, #COUNT(UInt8(1)) AS count_order
        Aggregate: groupBy=[[#l_returnflag, #l_linestatus]], aggr=[[SUM(#l_quantity), SUM(#l_extendedprice), SUM(#l_extendedprice Multiply Int64(1) Minus #l_discount), SUM(#l_extendedprice Multiply Int64(1) Minus #l_discount Multiply Int64(1) Plus #l_tax), AVG(#l_quantity), AVG(#l_extendedprice), AVG(#l_discount), COUNT(UInt8(1))]]
          Filter: #l_shipdate LtEq CAST(Utf8("1998-09-02") AS Date32(Day))
            TableScan: lineitem projection=Some([4, 5, 6, 7, 8, 9, 10])
Query 1 iteration 9 took 802.7 ms
Query 1 avg time: 915.30 ms

[codecov-io]

Codecov Report

Merging #8986 (aacc33d) into master (38ba81b) will decrease coverage by 0.00%.
The diff coverage is 66.66%.

@@            Coverage Diff             @@
##           master    #8986      +/-   ##
==========================================
- Coverage   82.65%   82.65%   -0.01%     
==========================================
  Files         200      200              
  Lines       49675    49678       +3     
==========================================
+ Hits        41058    41060       +2     
- Misses       8617     8618       +1     

Impacted Files	Coverage Δ
rust/arrow-flight/src/arrow.flight.protocol.rs	0.00% <ø> (ø)
rust/benchmarks/src/bin/tpch.rs	0.00% <0.00%> (ø)
rust/datafusion/src/execution/context.rs	89.45% <100.00%> (+0.02%)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 38ba81b...aacc33d. Read the comment docs.

[github-actions]

https://issues.apache.org/jira/browse/ARROW-5950

[jorgecarleitao]

Cool. Yes, good idea.

FWIW for me we can add it to the bin, also.

[Dandandan]

FWIW for me we can add it to the bin, also.

I think that's a good idea, but don't want to add a logger as dependency directly to datafusion. Doing so at the bin level is not yet supported rust-lang/cargo#1982 It will probably change though rust-lang/rfcs#3020
Maybe at some point they can move to their own project.

[alamb]

I agree the idea of allowing logging, as well as the implementation of using the standard logging system is a good idea. Thanks @Dandandan!

It looks like a rebase is needed but otherwise I think this PR is ready to go

[Dandandan]

Thanks @alamb updated the PR