The cache command caches the entire table, with each column stored in its own byte buffer.  When querying the data, only the columns that you are asking for are scanned in memory.  I'm not sure what mechanism spark is using to report the amount of data read.

If you want to read only the data that you are looking for off of the disk, I'd suggest looking at parquet.

On Wed, Jan 7, 2015 at 1:37 AM, Xuelin Cao <xuelincao@yahoo.com.invalid> wrote:

Hi, 

      Curious and curious. I'm puzzled by the Spark SQL cached table.

      Theoretically, the cached table should be columnar table, and only scan the column that included in my SQL.

      However, in my test, I always see the whole table is scanned even though I only "select" one column in my SQL.

      Here is my code:

val sqlContext = new org.apache.spark.sql.SQLContext(sc)
import sqlContext._
sqlContext.jsonFile("/data/ad.json").registerTempTable("adTable")
sqlContext.cacheTable("adTable")  //The table has > 10 columns

//First run, cache the table into memory
sqlContext.sql("select * from adTable").collect

//Second run, only one column is used. It should only scan a small fraction of data
sqlContext.sql("select adId from adTable").collect 
sqlContext.sql("select adId from adTable").collect
sqlContext.sql("select adId from adTable").collect

        What I found is, every time I run the SQL, in WEB UI, it shows the total amount of input data is always the same --- the total amount of the table.

        Is anything wrong? My expectation is:
        1. The cached table is stored as columnar table
        2. Since I only need one column in my SQL, the total amount of input data showed in WEB UI should be very small

        But what I found is totally not the case. Why?

        Thanks