Class: Blazer::Adapters::SqlAdapter

Inherits:
BaseAdapter show all
Defined in:
lib/blazer/adapters/sql_adapter.rb

Direct Known Subclasses

SnowflakeAdapter

Instance Attribute Summary collapse

Attributes inherited from BaseAdapter

#data_source

Instance Method Summary collapse

Constructor Details

#initialize(data_source) ⇒ SqlAdapter

Returns a new instance of SqlAdapter.



6
7
8
9
10
11
12
13
14
15
16
 
# File 'lib/blazer/adapters/sql_adapter.rb', line 6

def initialize(data_source)
  super

  @connection_model =
    Class.new(Blazer::Connection) do
      def self.name
        "Blazer::Connection::Adapter#{object_id}"
      end
      establish_connection(data_source.settings["url"]) if data_source.settings["url"]
    end
end

Instance Attribute Details

#connection_model ⇒ Object (readonly)

Returns the value of attribute connection_model.



4
5
6
 
# File 'lib/blazer/adapters/sql_adapter.rb', line 4

def connection_model
  @connection_model
end

Instance Method Details

#cachable?(statement) ⇒ Boolean

Returns:

  • (Boolean) 


121
122
123
 
# File 'lib/blazer/adapters/sql_adapter.rb', line 121

def cachable?(statement)
  !%w[CREATE ALTER UPDATE INSERT DELETE].include?(statement.split.first.to_s.upcase)
end

#cancel(run_id) ⇒ Object 



110
111
112
113
114
115
116
117
118
119
 
# File 'lib/blazer/adapters/sql_adapter.rb', line 110

def cancel(run_id)
  if postgresql?
    select_all("SELECT pg_cancel_backend(pid) FROM pg_stat_activity WHERE pid <> pg_backend_pid() AND query LIKE ?", ["%,run_id:#{run_id}%"])
  elsif redshift?
    first_row = select_all("SELECT pid FROM stv_recents WHERE status = 'Running' AND query LIKE ?", ["%,run_id:#{run_id}%"]).first
    if first_row
      select_all("CANCEL #{first_row["pid"].to_i}")
    end
  end
end

#cohort_analysis_statement(statement, period:, days:) ⇒ Object

TODO treat date columns as already in time zone



130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
 
# File 'lib/blazer/adapters/sql_adapter.rb', line 130

def cohort_analysis_statement(statement, period:, days:)
  raise "Cohort analysis not supported" unless supports_cohort_analysis?

  cohort_column = statement =~ /\bcohort_time\b/ ? "cohort_time" : "conversion_time"
  tzname = Blazer.time_zone.tzinfo.name

  if mysql?
    time_sql = "CONVERT_TZ(cohorts.cohort_time, '+00:00', ?)"
    case period
    when "day"
      date_sql = "CAST(DATE_FORMAT(#{time_sql}, '%Y-%m-%d') AS DATE)"
      date_params = [tzname]
    when "week"
      date_sql = "CAST(DATE_FORMAT(#{time_sql} - INTERVAL ((5 + DAYOFWEEK(#{time_sql})) % 7) DAY, '%Y-%m-%d') AS DATE)"
      date_params = [tzname, tzname]
    else
      date_sql = "CAST(DATE_FORMAT(#{time_sql}, '%Y-%m-01') AS DATE)"
      date_params = [tzname]
    end
    bucket_sql = "CAST(CEIL(TIMESTAMPDIFF(SECOND, cohorts.cohort_time, query.conversion_time) / ?) AS INTEGER)"
  else
    date_sql = "date_trunc(?, cohorts.cohort_time::timestamptz AT TIME ZONE ?)::date"
    date_params = [period, tzname]
    bucket_sql = "CEIL(EXTRACT(EPOCH FROM query.conversion_time - cohorts.cohort_time) / ?)::int"
  end

  # WITH not an optimization fence in Postgres 12+
  statement = <<~SQL
    WITH query AS (
      #{statement}
    ),
    cohorts AS (
      SELECT user_id, MIN(#{cohort_column}) AS cohort_time FROM query
      WHERE user_id IS NOT NULL AND #{cohort_column} IS NOT NULL
      GROUP BY 1
    )
    SELECT
      #{date_sql} AS period,
      0 AS bucket,
      COUNT(DISTINCT cohorts.user_id)
    FROM cohorts GROUP BY 1
    UNION ALL
    SELECT
      #{date_sql} AS period,
      #{bucket_sql} AS bucket,
      COUNT(DISTINCT query.user_id)
    FROM cohorts INNER JOIN query ON query.user_id = cohorts.user_id
    WHERE query.conversion_time IS NOT NULL
    AND query.conversion_time >= cohorts.cohort_time
    #{cohort_column == "conversion_time" ? "AND query.conversion_time != cohorts.cohort_time" : ""}
    GROUP BY 1, 2
  SQL
  params = [statement] + date_params + date_params + [days.to_i * 86400]
  connection_model.send(:sanitize_sql_array, params)
end

#cost(statement) ⇒ Object 



84
85
86
87
88
89
90
91
92
 
# File 'lib/blazer/adapters/sql_adapter.rb', line 84

def cost(statement)
  result = explain(statement)
  if sqlserver?
    result["TotalSubtreeCost"]
  else
    match = /cost=\d+\.\d+..(\d+\.\d+) /.match(result)
    match[1] if match
  end
end

#explain(statement) ⇒ Object 



94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
 
# File 'lib/blazer/adapters/sql_adapter.rb', line 94

def explain(statement)
  if postgresql? || redshift?
    select_all("EXPLAIN #{statement}").rows.first.first
  elsif sqlserver?
    begin
      execute("SET SHOWPLAN_ALL ON")
      result = select_all(statement).each.first
    ensure
      execute("SET SHOWPLAN_ALL OFF")
    end
    result
  end
rescue
  nil
end

#preview_statement ⇒ Object 



72
73
74
75
76
77
78
 
# File 'lib/blazer/adapters/sql_adapter.rb', line 72

def preview_statement
  if sqlserver?
    "SELECT TOP (10) * FROM {table}"
  else
    "SELECT * FROM {table} LIMIT 10"
  end
end

#reconnect ⇒ Object 



80
81
82
 
# File 'lib/blazer/adapters/sql_adapter.rb', line 80

def reconnect
  connection_model.establish_connection(settings["url"])
end

#run_statement(statement, comment) ⇒ Object 



18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
 
# File 'lib/blazer/adapters/sql_adapter.rb', line 18

def run_statement(statement, comment)
  columns = []
  rows = []
  error = nil

  begin
    in_transaction do
      set_timeout(data_source.timeout) if data_source.timeout

      result = select_all("#{statement} /*#{comment}*/")
      columns = result.columns
      result.rows.each do |untyped_row|
        rows << (result.column_types.empty? ? untyped_row : columns.each_with_index.map { |c, i| untyped_row[i] && result.column_types[c] ? result.column_types[c].send(:cast_value, untyped_row[i]) : untyped_row[i] })
      end
    end
  rescue => e
    error = e.message.sub(/.+ERROR: /, "")
    error = Blazer::TIMEOUT_MESSAGE if Blazer::TIMEOUT_ERRORS.any? { |e| error.include?(e) }
    reconnect if error.include?("PG::ConnectionBad")
  end

  [columns, rows, error]
end

#schema ⇒ Object 



66
67
68
69
70
 
# File 'lib/blazer/adapters/sql_adapter.rb', line 66

def schema
  sql = add_schemas("SELECT table_schema, table_name, column_name, data_type, ordinal_position FROM information_schema.columns")
  result = data_source.run_statement(sql)
  result.rows.group_by { |r| [r[0], r[1]] }.map { |k, vs| {schema: k[0], table: k[1], columns: vs.sort_by { |v| v[2] }.map { |v| {name: v[2], data_type: v[3]} }} }.sort_by { |t| [t[:schema] == default_schema ? "" : t[:schema], t[:table]] }
end

#supports_cohort_analysis? ⇒ Boolean

Returns:

  • (Boolean) 


125
126
127
 
# File 'lib/blazer/adapters/sql_adapter.rb', line 125

def supports_cohort_analysis?
  postgresql? || mysql?
end

#tables ⇒ Object 



42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
 
# File 'lib/blazer/adapters/sql_adapter.rb', line 42

def tables
  sql = add_schemas("SELECT table_schema, table_name FROM information_schema.tables")
  result = data_source.run_statement(sql, refresh_cache: true)
  if postgresql? || redshift? || snowflake?
    result.rows.sort_by { |r| [r[0] == default_schema ? "" : r[0], r[1]] }.map do |row|
      table =
        if row[0] == default_schema
          row[1]
        else
          "#{row[0]}.#{row[1]}"
        end

      table = table.downcase if snowflake?

      {
        table: table,
        value: connection_model.connection.quote_table_name(table)
      }
    end
  else
    result.rows.map(&:second).sort
  end
end