From 4fbb071e3ef260c34d883664a6d8627ce9547452 Mon Sep 17 00:00:00 2001
From: Stas <stanislav@perfectline.co>
Date: Thu, 3 Mar 2016 19:39:07 +0200
Subject: [PATCH 1/3] 113146419-new_task

---
 lib/tasks/epp.rake | 31 ++++++++++++++++++++++++++-----
 1 file changed, 26 insertions(+), 5 deletions(-)

diff --git a/lib/tasks/epp.rake b/lib/tasks/epp.rake
index e03835dad..99062994f 100644
--- a/lib/tasks/epp.rake
+++ b/lib/tasks/epp.rake
@@ -8,12 +8,33 @@ namespace :epp do
   desc 'Trim logs'
   task trim_documents: :environment do
     puts '-----> Running query'
-    sql = <<-SQL
-    UPDATE epp_logs SET request = regexp_replace(request, '<eis:legalDocument(.|\n)*?<\/eis:legalDocument>', '<eis:legalDocument>[FILTERED]<\eis:legalDocument>');
-    SQL
-    ApiLog::EppLog.connection.execute(sql)
+    puts '-----> Selecting count of all rows'
 
-    puts "-----> Query done"
+    rows = ApiLog::EppLog.where("request ~* ?", '<eis:legalDocument(.|\n)*?<\/eis:legalDocument>')
+    count =  rows.count
+
+    puts "-----> Total rows #{count}"
+
+
+    i = 0
+    ids = []
+
+    ApiLog::EppLog.where("request ~* ?", '<eis:legalDocument(.|\n)*?<\/eis:legalDocument>').find_each(batch_size: 10000)do |x|
+
+    trimmed = x.request.gsub(/<eis:legalDocument([^>]+)>([^<])+<\/eis:legalDocument>/, "<eis:legalDocument>[FILTERED]</eis:legalDocument>")
+    x.request = trimmed
+    x.save
+
+    ids.push x.id
+    i += 1
+
+    if i = 500
+    i = 0
+      puts "500 rows updated #{ids.join(', ')}"
+    end
+
+    end
+  puts "-----> Query done"
   end
 end
 

From 334e2b1362ecb7e6230c140b0e2bdd45ac232938 Mon Sep 17 00:00:00 2001
From: Stas <stanislav@perfectline.co>
Date: Fri, 4 Mar 2016 17:20:34 +0200
Subject: [PATCH 2/3] 113146419-script_changes

---
 lib/tasks/epp.rake | 40 +++++++++++++++++++++++-----------------
 1 file changed, 23 insertions(+), 17 deletions(-)

diff --git a/lib/tasks/epp.rake b/lib/tasks/epp.rake
index 99062994f..361bbbc68 100644
--- a/lib/tasks/epp.rake
+++ b/lib/tasks/epp.rake
@@ -8,33 +8,39 @@ namespace :epp do
   desc 'Trim logs'
   task trim_documents: :environment do
     puts '-----> Running query'
-    puts '-----> Selecting count of all rows'
-
-    rows = ApiLog::EppLog.where("request ~* ?", '<eis:legalDocument(.|\n)*?<\/eis:legalDocument>')
-    count =  rows.count
-
-    puts "-----> Total rows #{count}"
 
+    start = Time.zone.now.to_f
 
     i = 0
     ids = []
+    count = 0
 
-    ApiLog::EppLog.where("request ~* ?", '<eis:legalDocument(.|\n)*?<\/eis:legalDocument>').find_each(batch_size: 10000)do |x|
+    ApiLog::EppLog
+         .where("request ~* ?", '<eis:legalDocument(.|\n)*?<\/eis:legalDocument>')
+         .where("request NOT LIKE ?", "%<eis:legalDocument>[FILTERED]</eis:legalDocument>%")
+         .where("request NOT LIKE ?", '%<eis:legalDocument type="pdf"></eis:legalDocument>%').find_each(batch_size: 1000)do |x|
 
-    trimmed = x.request.gsub(/<eis:legalDocument([^>]+)>([^<])+<\/eis:legalDocument>/, "<eis:legalDocument>[FILTERED]</eis:legalDocument>")
-    x.request = trimmed
-    x.save
+      trimmed = x.request.gsub(/<eis:legalDocument([^>]+)>([^<])+<\/eis:legalDocument>/, "<eis:legalDocument>[FILTERED]</eis:legalDocument>")
 
-    ids.push x.id
-    i += 1
+      x.request = trimmed
 
-    if i = 500
-    i = 0
-      puts "500 rows updated #{ids.join(', ')}"
-    end
+      x.save and ids.push x.id and i += 1 and ids.push x.id
+
+      i += 1
+      count += 1
+
+      if i == 100
+        puts "-----> Total rows updated #{count}"
+        puts "Last #{i} rows ids #{ids.join(', ')}"
+        i = 0
+        ids = []
+      end
 
     end
-  puts "-----> Query done"
+
+  puts "-----> Total rows updated #{count}"
+  puts "Last #{count} rows ids #{ids.join(', ')}"
+  puts "-----> Query done total #{(Time.zone.now.to_f - start).round(2)} seconds"
   end
 end
 

From ce06a95274521392ebd9b5beed5c56979950c717 Mon Sep 17 00:00:00 2001
From: Stas <stanislav@perfectline.co>
Date: Mon, 7 Mar 2016 15:00:10 +0200
Subject: [PATCH 3/3] 113146419-step_size_changed

---
 lib/tasks/epp.rake | 7 ++-----
 1 file changed, 2 insertions(+), 5 deletions(-)

diff --git a/lib/tasks/epp.rake b/lib/tasks/epp.rake
index 361bbbc68..380fa3248 100644
--- a/lib/tasks/epp.rake
+++ b/lib/tasks/epp.rake
@@ -24,12 +24,9 @@ namespace :epp do
 
       x.request = trimmed
 
-      x.save and ids.push x.id and i += 1 and ids.push x.id
+      x.save and count += 1 and i += 1 and ids.push x.id
 
-      i += 1
-      count += 1
-
-      if i == 100
+      if i == 500
         puts "-----> Total rows updated #{count}"
         puts "Last #{i} rows ids #{ids.join(', ')}"
         i = 0