From 1962175cc66502848aa0e17116959c3e0986f140 Mon Sep 17 00:00:00 2001 From: f Date: Sat, 30 Apr 2022 10:58:41 -0300 Subject: [PATCH 1/2] Revert "mostrar demora" This reverts commit e9d13a7c5931b9ad4e86bf218ff6aff79ac132c0. --- app/jobs/uri_collection_job.rb | 22 ---------------------- 1 file changed, 22 deletions(-) diff --git a/app/jobs/uri_collection_job.rb b/app/jobs/uri_collection_job.rb index ae4b098e..196d2e50 100644 --- a/app/jobs/uri_collection_job.rb +++ b/app/jobs/uri_collection_job.rb @@ -38,10 +38,7 @@ class UriCollectionJob < PeriodicJob host_dimensions = { host: host } columns.each_key do |column| - time_start = Time.now columns[column] = AccessLog.where(**host_dimensions).distinct(column).pluck(column) - - puts "#{column}: #{Time.now - time_start}" end # Las URIs son la fuente de verdad de las visitas, porque son las @@ -54,58 +51,39 @@ class UriCollectionJob < PeriodicJob name = 'host|uri' dimensions = { host: host, uri: uri } - time_start = Time.now rollup(name, beginning, **dimensions) - puts "#{dimensions.values.join}: #{Time.now - time_start}" - - time_start = Time.now reduce_rollup(name, beginning, **dimensions) - puts "reduced: #{Time.now - time_start}" columns.each_pair do |column, values| # Obtener orígenes de visitas por host values.each do |value| - time_start = Time.now - column_name = "host|uri|#{column}" dimensions[column] = value rollup(column_name, beginning, **dimensions) - puts "#{dimensions.values.join}: #{Time.now - time_start}" - - time_start = Time.now reduce_rollup(column_name, beginning, **dimensions) - puts "reduced: #{Time.now - time_start}" end end end # Reducir todas las visitas a cantidad de visitas por host - time_start = Time.now square_rollup(name: 'host|uri', new_name: 'host', interval: starting_interval, dimensions: host_dimensions, beginning: beginning) - puts "#{host}: #{Time.now - time_start}" # Acumular por mes y año - time_start = Time.now reduce_rollup('host', beginning, **host_dimensions) - puts "reduced: #{Time.now - time_start}" columns.each_key do |column| - time_start = Time.now square_rollup(name: "host|uri|#{column}", new_name: "host|#{column}", interval: starting_interval, dimensions: host_dimensions, beginning: beginning) - puts "#{host} #{column}: #{Time.now - time_start}" - time_start = Time.now reduce_rollup("host|#{column}", beginning, **host_dimensions) - puts "reduced: #{Time.now - time_start}" end end From 6699db838078d141cc3eaab7815b09d22d1862ec Mon Sep 17 00:00:00 2001 From: f Date: Sat, 30 Apr 2022 11:34:21 -0300 Subject: [PATCH 2/2] otra vez el mismo error de hashes acumulados hice benchmarks, generar hashes nuevos es mas rapido que duplicarlos, y eliminar las keys que no corresponden es apenas mas rapido pero con mas posibilidad de bugs --- app/jobs/uri_collection_job.rb | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/app/jobs/uri_collection_job.rb b/app/jobs/uri_collection_job.rb index 196d2e50..620123db 100644 --- a/app/jobs/uri_collection_job.rb +++ b/app/jobs/uri_collection_job.rb @@ -58,10 +58,11 @@ class UriCollectionJob < PeriodicJob # Obtener orígenes de visitas por host values.each do |value| column_name = "host|uri|#{column}" - dimensions[column] = value + column_dimensions = { host: host, uri: uri } + column_dimensions[column] = value - rollup(column_name, beginning, **dimensions) - reduce_rollup(column_name, beginning, **dimensions) + rollup(column_name, beginning, **column_dimensions) + reduce_rollup(column_name, beginning, **column_dimensions) end end end