varchar columns are empty when reading from a Trino source

[jhatcher1]

Issue Description

• Description of the issue:

  I am reading data from a Trino database, and saving it to both CSV and parquet files. In both cases, the sling CLI generates the files without errors.

  Upon inspecting the resulting CSV and parquet files, any columns of type varchar or varchar(n) are blank in the resulting file.

  I was able to reroduce the issue running Trino locally creating a simple table:

  trino> CREATE SCHEMA iceberg.example_schema;
  CREATE SCHEMA
  trino> CREATE TABLE iceberg.example_schema.example_sling (name varchar, id bigint, description varchar);
  CREATE TABLE
  trino> INSERT INTO iceberg.example_schema.example_sling VALUES ('my_name', 123456, 'my_description');
  INSERT: 1 row
  trino> SELECT * FROM iceberg.example_schema.example_sling;
    name   |   id   |  description
  ---------+--------+----------------
   my_name | 123456 | my_description
  (1 row)
  

  I then ran the following to export the table to CSV:

  sling conns set LOCAL_TRINO type=trino http_url="http://admin@localhost:8080?catalog=iceberg&schema=example_schema"
  sling run --src-conn LOCAL_TRINO --src-stream "example_schema.example_sling" --tgt-object "file://example_sling.csv"
  

  Inspecting the CSV I see:

  $ cat example_sling.csv
  name,id,description
  ,123456,
  

  I observe the same when saving to a parquet file, and open it in dbeaver (via duckdb):

  

  I tried using both an iceberg catalog and a postgres catalog when creating the table to determine if the issue was catalog specific, but I saw the same behaviour regardless of the catalog used.

• Sling version (sling --version): 1.2.10

• Operating System (linux, mac, windows): mac (arm)

• Log Output (please run command with -d):

$ sling conns set LOCAL_TRINO_POSTGRES type=trino http_url="http://admin@localhost:8080?catalog=postgres&schema=public"
$ sling run --src-conn LOCAL_TRINO_POSTGRES --src-stream "postgres.public.example_sling" --tgt-object "file://example_sling_pg.csv" -d
2024-05-23 15:10:43 DBG Sling version: 1.2.10 (darwin arm64)
2024-05-23 15:10:43 DBG type is db-file
2024-05-23 15:10:43 DBG using source options: {"empty_as_null":false,"null_if":"NULL","datetime_format":"AUTO","max_decimals":-1}
2024-05-23 15:10:43 DBG using target options: {"header":true,"compression":"auto","concurrency":7,"datetime_format":"auto","delimiter":",","file_max_rows":0,"file_max_bytes":0,"max_decimals":-1,"use_bulk":true,"add_new_columns":true,"column_casing":"source"}
2024-05-23 15:10:43 INF connecting to source database (trino)
2024-05-23 15:10:43 DBG opened "trino" connection (conn-trino-Xqx)
2024-05-23 15:10:43 INF reading from source database
2024-05-23 15:10:43 DBG select * from "postgres"."public"."example_sling"
2024-05-23 15:10:43 INF writing to target file system (file)
2024-05-23 15:10:43 DBG writing to file://example_sling_pg.csv [fileRowLimit=0 fileBytesLimit=0 compression=auto concurrency=7 useBufferedStream=false fileFormat=csv]
2024-05-23 15:10:43 DBG wrote 30 B: 1 rows [6 r/s]
2024-05-23 15:10:43 INF wrote 1 rows [6 r/s] to file://example_sling_pg.csv
2024-05-23 15:10:43 DBG closed "trino" connection (conn-trino-Xqx)
2024-05-23 15:10:43 INF execution succeeded