active_hll 0.1.0 → 0.1.1

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA256:
-  metadata.gz: a9014f39ae62f6db1066d4933bc28485ecc57e788922d70437c4744edfb0c033
-  data.tar.gz: 91320d88909c41e425966993e145b9ff8aa5b57c6c3ac0556dc3661519691968
+  metadata.gz: 20a8025b26b6bc13e6ff8a37b877130200b87ca60608a68d1d96ba167c8c1a76
+  data.tar.gz: 4f3d16a8da8a25554ee8d7f9e5d1d7e6d6621e083528fde848e18b88daa0bbbf
 SHA512:
-  metadata.gz: c55980654230a259a249a24b959c18b054d8e41773653bcc5411dbbbe3200fdd35c65dcfe729ed042943a8f2e81e26843c696d4989840fa43384622e3d56614f
-  data.tar.gz: 5285ca0005787997dec664512b0bc6796201a8a4bd719e0ce0b9f6efa50096f3a3e7ade096db6ba943e9f84630b2a92c5d9b3a53a4a0b744cc4162e87d6c2990
+  metadata.gz: 856f6b94a988a75b4809b375f9f554c781f07958a89015c6b4cfa6eaf0c67bfaf12896762f1d940ef0c78f1bf26d674fb02a4f7bcf078205fbbd7a033a7bace0
+  data.tar.gz: c629cee7c89b0ee9b05ef65a453ee21be21f696bfd399ab7f29ae50f33aa369f6feeeb3ad41967e81329cadb1e70f78732e549c6541169e56e2f7f678fae7366

data/CHANGELOG.md

@@ -1,3 +1,7 @@
+## 0.1.1 (2023-01-29)
+
+- Added experimental `hll_upsert` method
+
 ## 0.1.0 (2023-01-24)
 
 - First release

data/README.md

@@ -53,7 +53,7 @@ You can use [batch](#batch) and [stream](#stream) approaches to build HLLs
 
 ### Batch
 
-Use the `hll_agg` method to generate HLLs from existing data
+To generate HLLs from existing data, use the `hll_agg` method
 
 ```ruby
 hlls = Event.group_by_day(:created_at).hll_agg(:visitor_id)
@@ -74,12 +74,18 @@ For a large number of HLLs, use SQL to generate and upsert in a single statement
 
 ### Stream
 
-Use the `hll_add` method to add new data to HLLs
+To add new data to HLLs, use the `hll_add` method
 
 ```ruby
 EventRollup.where(time_bucket: Date.current).hll_add(visitor_ids: ["visitor1", "visitor2"])
 ```
 
+or the `hll_upsert` method (experimental)
+
+```ruby
+EventRollup.hll_upsert({time_bucket: Date.current, visitor_ids: ["visitor1", "visitor2"]})
+```
+
 ## Querying
 
 Get approximate unique values for a time range
@@ -132,7 +138,7 @@ There’s not a way to remove data from an HLL, so to delete data for a specific
 
 ## Hosted Postgres
 
-The `hll` extension is available on Amazon RDS, Google Cloud SQL, and DigitalOcean Managed Databases.
+The `hll` extension is available on a number of [hosted providers](https://github.com/ankane/active_hll/issues/4).
 
 ## History
 

data/lib/active_hll/model.rb

@@ -50,6 +50,36 @@ module ActiveHll
 
         update_all(set_clauses.join(", "))
       end
+
+      # experimental
+      def hll_upsert(attributes)
+        hll_columns, other_columns = attributes.keys.partition { |a| columns_hash[a.to_s]&.type == :hll }
+
+        # important! raise if column detection fails
+        if hll_columns.empty?
+          raise ArgumentError, "No hll columns"
+        end
+
+        quoted_table = connection.quote_table_name(table_name)
+
+        quoted_hll_columns = hll_columns.map { |k| connection.quote_column_name(k) }
+        quoted_other_columns = other_columns.map { |k| connection.quote_column_name(k) }
+        quoted_columns = quoted_other_columns + quoted_hll_columns
+
+        hll_values =
+          hll_columns.map do |k|
+            vs = attributes[k]
+            vs = [vs] unless vs.is_a?(Array)
+            vs.map { |v| Utils.hll_hash_sql(self, v) }.join(" || ")
+          end
+        other_values = other_columns.map { |k| connection.quote(attributes[k]) }
+
+        insert_values = other_values + hll_values.map { |v| "hll_empty()#{v.size > 0 ? " || #{v}" : ""}" }
+        update_values = quoted_hll_columns.zip(hll_values).map { |k, v| "#{k} = COALESCE(#{quoted_table}.#{k}, hll_empty())#{v.size > 0 ? " || #{v}" : ""}" }
+
+        sql = "INSERT INTO #{quoted_table} (#{quoted_columns.join(", ")}) VALUES (#{insert_values.join(", ")}) ON CONFLICT (#{quoted_other_columns.join(", ")}) DO UPDATE SET #{update_values.join(", ")}"
+        connection.exec_insert(sql, "#{name} Upsert")
+      end
     end
 
     # doesn't update in-memory record attribute for performance

data/lib/active_hll/version.rb

@@ -1,3 +1,3 @@
 module ActiveHll
-  VERSION = "0.1.0"
+  VERSION = "0.1.1"
 end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: active_hll
 version: !ruby/object:Gem::Version
-  version: 0.1.0
+  version: 0.1.1
 platform: ruby
 authors:
 - Andrew Kane
 autorequire:
 bindir: bin
 cert_chain: []
-date: 2023-01-24 00:00:00.000000000 Z
+date: 2023-01-30 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: activerecord