SAS EG & Hadoop Data Checks

skruz83 — Mon, 04 Jul 2016 15:14:44 GMT

Hi,

I am fairly new to hadoop and using SAS EG to access the data.

I want to run a series of data checks on the data that is stored in hadoop i.e. for a particular table (or library/database) for each (tables)columns identify the min, max, missing, no of records etc...

I tried using the traditional PROC Contents/ PROC Datasets but it takes ages given the volume of data etc..

Is there a better way to run the two commands in hadoop via hive sql?

Effectively I am after a table which shows:

table name, column_name, column type, no of records, no of missing values, no of distinct values, min value, max value, min length, max length,

Regards

Re: SAS EG & Hadoop Data Checks

LinusH — Tue, 05 Jul 2016 09:16:01 GMT

Proc contents should give the structure, but it shouldn't take much time.
The other stats isn't available in contents nor datasets procedure. For those i sugest that you use SQL. But it requires a full table scan given the nature of your requirement. Just be sure that the SQL is sent to Hive. Try with a small table first, and use
Options msglevel = I sastrace = ',,,d' SASTRACELOC = saslog;
for verification.

topic SAS EG & Hadoop Data Checks in SAS Data Management

SAS EG & Hadoop Data Checks

Re: SAS EG & Hadoop Data Checks