Hadoop - AWS EMR CLI - passer des Arguments à la ruche - Stack Overflow

I am using aws' emr ruby cli to generate Hadoop clusters and I am trying to include arguments to use within a HIVE script hosted elsewhere like so:

./elastic-mapreduce --create ... --args -d,DT=2013-01-26

'DT' shows up satisfactorily in my HadoopJarStep.Args array, so I try to include it in the HIVE script like so:

...

tblproperties(
  'dynamodb.table.name' = ${DT},
  ...

but I quickly get this:

Parse Error: line 8:28 mismatched input '$' expecting StringLiteral near '=' in specifying key/value property

How should I properly include the argument in my HIVE script?

I'm not exactly sure why your current approach isn't working, but I've been successful with:

./elastic-mapreduce --create ... --args "-hiveconf,DT=2013-01-26"

and in the hive script:

tblproperties(
    "dynamodb.table.name" = "${hiveconf:DT}",
    ...
)

Hope this helps.

I am using aws' emr ruby cli to generate Hadoop clusters and I am trying to include arguments to use within a HIVE script hosted elsewhere like so:

./elastic-mapreduce --create ... --args -d,DT=2013-01-26

'DT' shows up satisfactorily in my HadoopJarStep.Args array, so I try to include it in the HIVE script like so:

...

tblproperties(
  'dynamodb.table.name' = ${DT},
  ...

but I quickly get this:

Parse Error: line 8:28 mismatched input '$' expecting StringLiteral near '=' in specifying key/value property

How should I properly include the argument in my HIVE script?

I'm not exactly sure why your current approach isn't working, but I've been successful with:

./elastic-mapreduce --create ... --args "-hiveconf,DT=2013-01-26"

and in the hive script:

tblproperties(
    "dynamodb.table.name" = "${hiveconf:DT}",
    ...
)

Hope this helps.

Source