将 json.gz 的命令从 s3 复制到 redshift

Question

We have stored results of DDB to s3 output as json.gz files. We want to transfer those to redshift using the copy command. We don't want to do a direct DDB to Redshift because copying directly usually involves a scan operation. This causes read capacity to be utilized which we want to avoid since these tables are pretty large. I could not find much on how to use a copy command on a json.gz file. Please let me know if someone can find a way to do this. I tried treating it like a json as suggested in one of the comments
copy itemtable from 's3://bucket/path/file.json.gz' iam_role '<role>' json 'auto ignorecase'
it did not work. My file when unzipped is in this format:
{"Item":{"field":{"S":"value"},"field":{"N":"value"}}}'\n'{"Item":{"field":{"S":"value"},"field":{"N":"value"}}}'\n'
Exact Error is
error is Load into table 'itemtable' failed. Check 'stl_load_errors' system table for details

Answer 1

Doing a few things worked for me.

from '<input s3 location>'
iam_role '<iam role>'
json '<jsonpath file location>' gzip ACCEPTINVCHARS ' ' TRUNCATECOLUMNS TRIMBLANKS
region '<aws region>'

here jsonpath.json file is in following format

{
  "jsonpaths": [
    "$['Item']['Field1']['S']",
    "$['Item']['Field2']['N']",
    .
    .
    .
  ]
}

And the table contains the same columns as fields specified in jsonpath.

As suggested in John Rotenstein's comment, the copy command takes care of gzip and we don't need to worry about it.