API Latest Releases: Java Core, C++ Core, Python, Memory, Pig, Hive,

CPC Sketch Pig UDFs


  • get jars
  • save the following script as cpc.pig
  • adjust jar versions and paths if necessary
  • save the below data into a file called “data.txt”
  • copy data to hdfs: “hdfs dfs -copyFromLocal data.txt”
  • run pig script: “pig cpc.pig”

cpc.pig script: building sketches, merging sketches and getting estimates

register datasketches-memory-2.0.0.jar;
register datasketches-java-3.1.0.jar;
register datasketches-pig-1.1.0.jar;

define dataToSketch org.apache.datasketches.pig.cpc.DataToSketch('12');       
define unionSketch org.apache.datasketches.pig.cpc.UnionSketch('12');       
define getEstimate org.apache.datasketches.pig.cpc.GetEstimate();     
define getEstimateAndBounds org.apache.datasketches.pig.cpc.GetEstimateAndErrorBounds('3');       
define toString org.apache.datasketches.pig.cpc.SketchToString();              

a = load 'data.txt' as (id, category);
b = group a by category;
c = foreach b generate flatten(group) as (category), dataToSketch(a.id) as sketch;
-- Sketches can be stored at this point in binary format to be used later:
-- store c into 'intermediate/$date' using BinStorage();
-- The next two lines print the results in human readable form for the purpose of this example
d = foreach c generate category, getEstimate(sketch);
dump d;

-- This can be a separate query
-- For example, the first part can produce a daily intermediate feed and store it,
-- and this part can load several instances of this daily intermediate feed and merge them
e = group c all;
f = foreach e generate unionSketch(c.sketch) as sketch;
g = foreach f generate getEstimate(sketch);
dump g;

h = foreach f generate flatten(getEstimateAndBounds(sketch)) as (estimate, lb, ub);
dump h;

data.txt (tab separated)

The example input data has 2 fields: id and category. There are 2 categories ‘a’ and ‘b’ with 50 unique IDs in each. Most of the IDs in these categories overlap, so that there are 60 unique IDs in total.

Results: From ‘dump d’:


From ‘dump g’ (merged across categories):


From ‘dump h’ (with error bounds, 99% confidence interval):