[parquet-cli] dictionary command throw NPE when specified column isn't dictionary encoding

'dictionary' command of parquet-cli throw NPE when specified column isn't dictionary encoding.
```Java

$ java -cp 'target/classes:target/dependency/*' org.apache.parquet.cli.Main dictionary /work/parquet-mr/data/test.parquet -c binary_field
Unknown error
java.lang.NullPointerException
at org.apache.parquet.cli.commands.ShowDictionaryCommand.run(ShowDictionaryCommand.java:78)
at org.apache.parquet.cli.Main.run(Main.java:147)
at org.apache.hadoop.util.ToolRunner.run(ToolRunner.java:70)
at org.apache.parquet.cli.Main.main(Main.java:177)
```

The schema of 'test.parquet' is following:
```Java

$ java -cp 'target/classes:target/dependency/*' org.apache.parquet.cli.Main meta /work/parquet-mr/data/test.parquet

File path: /work/parquet-mr/data/test.parquet
Created by: parquet-mr version 1.12.0-SNAPSHOT (build 1e62e2e2ca903d4109480bc87ceec1dc954b6c92)
Properties:
writer.model.name: example
Schema:
message test {
required int32 int32_field;
required int64 int64_field;
required float float_field;
required double double_field;
required binary binary_field;
required int64 timestamp_field (TIMESTAMP(MILLIS,true));
}

Row group 0: count: 395 15.87 B records start: 4 total: 6.120 kB
--------------------------------------------------------------------------------
type encodings count avg size nulls min / max
int32_field INT32 _ D 395 0.20 B 0 "32" / "426"
int64_field INT64 _ D 395 0.20 B 0 "64" / "458"
float_field FLOAT _ _ 395 4.13 B 0 "1.0" / "395.0"
double_field DOUBLE _ _ 395 8.13 B 0 "2.0" / "396.0"
binary_field BINARY _ D 395 2.98 B 0 "0x6162636465666768696A6B6..." / "0x6162636465666768696A6B6..."
timestamp_field INT64 _ D 395 0.23 B 0 "2018-11-04T12:41:15.123+0000" / "2018-11-04T12:47:49.123+0000"

Row group 1: count: 395 15.92 B records start: 6271 total: 6.142 kB
--------------------------------------------------------------------------------
type encodings count avg size nulls min / max
int32_field INT32 _ D 395 0.20 B 0 "427" / "821"
int64_field INT64 _ D 395 0.20 B 0 "459" / "853"
float_field FLOAT _ _ 395 4.13 B 0 "396.0" / "790.0"
double_field DOUBLE _ _ 395 8.13 B 0 "397.0" / "791.0"
binary_field BINARY _ D 395 3.03 B 0 "0x6162636465666768696A6B6..." / "0x6162636465666768696A6B6..."
timestamp_field INT64 _ D 395 0.23 B 0 "2018-11-04T12:47:50.123+0000" / "2018-11-04T12:54:24.123+0000"

Row group 2: count: 234 16.53 B records start: 12560 total: 3.777 kB
--------------------------------------------------------------------------------
type encodings count avg size nulls min / max
int32_field INT32 _ D 234 0.17 B 0 "822" / "1055"
int64_field INT64 _ D 234 0.31 B 0 "854" / "1087"
float_field FLOAT _ _ 234 4.11 B 0 "791.0" / "1024.0"
double_field DOUBLE _ _ 234 8.21 B 0 "792.0" / "1025.0"
binary_field BINARY _ D 234 3.38 B 0 "0x6162636465666768696A6B6..." / "0x6162636465666768696A6B6..."
timestamp_field INT64 _ D 234 0.35 B 0 "2018-11-04T12:54:25.123+0000" / "2018-11-04T12:58:18.123+0000"
```

**Reporter**: [Masayuki Takahashi](https://issues.apache.org/jira/secure/ViewProfile.jspa?name=masayuki038) / @masayuki038
#### Original Issue Attachments:
- [test.parquet](https://issues.apache.org/jira/secure/attachment/12959433/test.parquet)

_{**Note**: *This issue was originally created as [PARQUET-1535](https://issues.apache.org/jira/browse/PARQUET-1535). Please see the [migration documentation](https://issues.apache.org/jira/browse/PARQUET-2502) for further details.*}

Reproduce the failure with parquet-cli's dictionary command against the attached test.parquet, selecting float_field, which is not dictionary encoded. Inspect ShowDictionaryCommand.java at line 78 and the command entry point to identify the null value. Done means a non-dictionary column no longer causes an NPE and the command reports the condition appropriately.

Written by the indexing model from the issue text.