You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@datasketches.apache.org by al...@apache.org on 2025/02/12 03:03:54 UTC
(datasketches-bigquery) 01/01: tuple readme
This is an automated email from the ASF dual-hosted git repository.
alsay pushed a commit to branch tuple_readme
in repository https://gitbox.apache.org/repos/asf/datasketches-bigquery.git
commit f30334c612f246351c3a4610749545673dba56d6
Author: AlexanderSaydakov <Al...@users.noreply.github.com>
AuthorDate: Tue Feb 11 19:03:24 2025 -0800
tuple readme
---
tuple/README.md | 19 +++++++++++++++----
tuple/README_template.md | 19 +++++++++++++++----
2 files changed, 30 insertions(+), 8 deletions(-)
diff --git a/tuple/README.md b/tuple/README.md
index 66f0125..32f9a44 100644
--- a/tuple/README.md
+++ b/tuple/README.md
@@ -19,10 +19,21 @@
# Apache DataSketches Tuple Sketches for Google BigQuery
-Tuple sketches extend the functionality of Theta sketches by
-allowing you to associate a summary value with each item in the set. This
-enables calculations like the sum, minimum, or maximum of values associated with
-the distinct items.
+Tuple sketches extend the functionality of Theta sketches by adding a Summary object associated
+with each distinct key retained by the sketch. When the identifier of an input pair (identifier, value) matches a unique
+key of the sketch, the associated Summary of that key can be modified based on user-defined policy.
+The set of all Summary values collected by the sketch represents a uniform random sample over the unique identifiers
+subset of all identifiers. This enables the use of common statistical computations of the Summary values, which can be extrapolated to the entire
+set of unique identifiers.
+
+The underlying C++ library supports Summary objects of any type (including complex types) and arbitrary policies
+of updating Summaries during the sketch building process, and combining these Summaries during union and intersection set operations.
+
+The current set of functions for BigQuery implements Summary objects as INT64 (unsigned in C++) with SUM, MIN, MAX, ONE (constant 1) policies (modes).
+This enables calculations like the sum, average, minimum, or maximum of the Summary values associated with the distinct keys.
+
+This implementation can serve as an example of how to implement Tuple sketch with a Summary type and policy of your choice.
+We are open to suggestions on what Summary types and policies to consider for inclusion here.
Please visit
[Tuple Sketches](https://datasketches.apache.org/docs/Tuple/TupleSketches.html)
diff --git a/tuple/README_template.md b/tuple/README_template.md
index 5efa530..a3abc98 100644
--- a/tuple/README_template.md
+++ b/tuple/README_template.md
@@ -19,10 +19,21 @@
# Apache DataSketches Tuple Sketches for Google BigQuery
-Tuple sketches extend the functionality of Theta sketches by
-allowing you to associate a summary value with each item in the set. This
-enables calculations like the sum, minimum, or maximum of values associated with
-the distinct items.
+Tuple sketches extend the functionality of Theta sketches by adding a Summary object associated
+with each distinct key retained by the sketch. When the identifier of an input pair (identifier, value) matches a unique
+key of the sketch, the associated Summary of that key can be modified based on user-defined policy.
+The set of all Summary values collected by the sketch represents a uniform random sample over the unique identifiers
+subset of all identifiers. This enables the use of common statistical computations of the Summary values, which can be extrapolated to the entire
+set of unique identifiers.
+
+The underlying C++ library supports Summary objects of any type (including complex types) and arbitrary policies
+of updating Summaries during the sketch building process, and combining these Summaries during union and intersection set operations.
+
+The current set of functions for BigQuery implements Summary objects as INT64 (unsigned in C++) with SUM, MIN, MAX, ONE (constant 1) policies (modes).
+This enables calculations like the sum, average, minimum, or maximum of the Summary values associated with the distinct keys.
+
+This implementation can serve as an example of how to implement Tuple sketch with a Summary type and policy of your choice.
+We are open to suggestions on what Summary types and policies to consider for inclusion here.
Please visit
[Tuple Sketches](https://datasketches.apache.org/docs/Tuple/TupleSketches.html)
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscribe@datasketches.apache.org
For additional commands, e-mail: commits-help@datasketches.apache.org