Class: Fluent::BigQuery::Writer
- Inherits:
-
Object
- Object
- Fluent::BigQuery::Writer
- Defined in:
- lib/fluent/plugin/bigquery/writer.rb
Instance Method Summary collapse
- #client ⇒ Object
- #create_load_job(chunk_id, project, dataset, table_id, upload_source, fields) ⇒ Object
- #create_table(project, dataset, table_id, record_schema) ⇒ Object
- #fetch_schema(project, dataset, table_id) ⇒ Object
-
#initialize(log, auth_method, options = {}) ⇒ Writer
constructor
A new instance of Writer.
- #insert_rows(project, dataset, table_id, rows, template_suffix: nil) ⇒ Object
- #wait_load_job(chunk_id, project, dataset, job_id, table_id) ⇒ Object
Constructor Details
#initialize(log, auth_method, options = {}) ⇒ Writer
Returns a new instance of Writer.
4 5 6 7 8 9 10 11 12 |
# File 'lib/fluent/plugin/bigquery/writer.rb', line 4 def initialize(log, auth_method, = {}) @auth_method = auth_method @scope = "https://www.googleapis.com/auth/bigquery" @options = @log = log @num_errors_per_chunk = {} @cached_client_expiration = Time.now + 1800 end |
Instance Method Details
#client ⇒ Object
14 15 16 17 18 19 20 21 22 23 24 25 26 |
# File 'lib/fluent/plugin/bigquery/writer.rb', line 14 def client return @client if @client && @cached_client_expiration > Time.now client = Google::Apis::BigqueryV2::BigqueryService.new.tap do |cl| cl. = get_auth cl..open_timeout_sec = @options[:open_timeout_sec] if @options[:open_timeout_sec] cl..read_timeout_sec = @options[:timeout_sec] if @options[:timeout_sec] cl..send_timeout_sec = @options[:timeout_sec] if @options[:timeout_sec] end @cached_client_expiration = Time.now + 1800 @client = client end |
#create_load_job(chunk_id, project, dataset, table_id, upload_source, fields) ⇒ Object
128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 |
# File 'lib/fluent/plugin/bigquery/writer.rb', line 128 def create_load_job(chunk_id, project, dataset, table_id, upload_source, fields) configuration = { configuration: { load: { destination_table: { project_id: project, dataset_id: dataset, table_id: table_id, }, schema: { fields: fields.to_a, }, write_disposition: "WRITE_APPEND", source_format: source_format, ignore_unknown_values: @options[:ignore_unknown_values], max_bad_records: @options[:max_bad_records], } } } job_id = create_job_id(chunk_id, dataset, table_id, fields.to_a) if @options[:prevent_duplicate_load] configuration[:configuration][:load].merge!(create_disposition: "CREATE_NEVER") if @options[:time_partitioning_type] configuration.merge!({job_reference: {project_id: project, job_id: job_id}}) if job_id # If target table is already exist, omit schema configuration. # Because schema changing is easier. begin if client.get_table(project, dataset, table_id) configuration[:configuration][:load].delete(:schema) end rescue Google::Apis::ServerError, Google::Apis::ClientError, Google::Apis::AuthorizationError raise Fluent::BigQuery::UnRetryableError.new("Schema is empty") if fields.empty? end res = client.insert_job( project, configuration, { upload_source: upload_source, content_type: "application/octet-stream", } ) wait_load_job(chunk_id, project, dataset, res.job_reference.job_id, table_id) @num_errors_per_chunk.delete(chunk_id) rescue Google::Apis::ServerError, Google::Apis::ClientError, Google::Apis::AuthorizationError => e @client = nil reason = e.respond_to?(:reason) ? e.reason : nil log.error "job.load API", project_id: project, dataset: dataset, table: table_id, code: e.status_code, message: e., reason: reason if @options[:auto_create_table] && e.status_code == 404 && /Not Found: Table/i =~ e. # Table Not Found: Auto Create Table create_table( project, dataset, table_id, fields, ) raise "table created. send rows next time." end if job_id && e.status_code == 409 && e. =~ /Job/ # duplicate load job wait_load_job(chunk_id, project, dataset, job_id, table_id) @num_errors_per_chunk.delete(chunk_id) return end raise Fluent::BigQuery::Error.wrap(e) end |
#create_table(project, dataset, table_id, record_schema) ⇒ Object
28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 |
# File 'lib/fluent/plugin/bigquery/writer.rb', line 28 def create_table(project, dataset, table_id, record_schema) create_table_retry_limit = 3 create_table_retry_wait = 1 create_table_retry_count = 0 table_id = safe_table_id(table_id) begin definition = { table_reference: { table_id: table_id, }, schema: { fields: record_schema.to_a, } } if @options[:time_partitioning_type] definition[:time_partitioning] = { type: @options[:time_partitioning_type].to_s.upcase, expiration_ms: @options[:time_partitioning_expiration] ? @options[:time_partitioning_expiration] * 1000 : nil }.select { |_, value| !value.nil? } end client.insert_table(project, dataset, definition, {}) log.debug "create table", project_id: project, dataset: dataset, table: table_id @client = nil rescue Google::Apis::ServerError, Google::Apis::ClientError, Google::Apis::AuthorizationError => e @client = nil = e. if e.status_code == 409 && /Already Exists:/ =~ log.debug "already created table", project_id: project, dataset: dataset, table: table_id # ignore 'Already Exists' error return end reason = e.respond_to?(:reason) ? e.reason : nil log.error "tables.insert API", project_id: project, dataset: dataset, table: table_id, code: e.status_code, message: , reason: reason if Fluent::BigQuery::Error.retryable_error_reason?(reason) && create_table_retry_count < create_table_retry_limit sleep create_table_retry_wait create_table_retry_wait *= 2 create_table_retry_count += 1 retry else raise Fluent::BigQuery::UnRetryableError.new("failed to create table in bigquery", e) end end end |
#fetch_schema(project, dataset, table_id) ⇒ Object
77 78 79 80 81 82 83 84 85 86 87 88 |
# File 'lib/fluent/plugin/bigquery/writer.rb', line 77 def fetch_schema(project, dataset, table_id) res = client.get_table(project, dataset, table_id) schema = Fluent::BigQuery::Helper.deep_stringify_keys(res.schema.to_h[:fields]) log.debug "Load schema from BigQuery: #{project}:#{dataset}.#{table_id} #{schema}" schema rescue Google::Apis::ServerError, Google::Apis::ClientError, Google::Apis::AuthorizationError => e @client = nil = e. log.error "tables.get API", project_id: project, dataset: dataset, table: table_id, code: e.status_code, message: nil end |
#insert_rows(project, dataset, table_id, rows, template_suffix: nil) ⇒ Object
90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 |
# File 'lib/fluent/plugin/bigquery/writer.rb', line 90 def insert_rows(project, dataset, table_id, rows, template_suffix: nil) body = { rows: rows, skip_invalid_rows: @options[:skip_invalid_rows], ignore_unknown_values: @options[:ignore_unknown_values], } body.merge!(template_suffix: template_suffix) if template_suffix res = client.insert_all_table_data(project, dataset, table_id, body, {}) log.debug "insert rows", project_id: project, dataset: dataset, table: table_id, count: rows.size if res.insert_errors && !res.insert_errors.empty? log.warn "insert errors", project_id: project, dataset: dataset, table: table_id, insert_errors: res.insert_errors.to_s if @options[:allow_retry_insert_errors] is_included_any_retryable_insert_error = res.insert_errors.any? do |insert_error| insert_error.errors.any? { |error| Fluent::BigQuery::Error.retryable_insert_errors_reason?(error.reason) } end if is_included_any_retryable_insert_error raise Fluent::BigQuery::RetryableError.new("failed to insert into bigquery(insert errors), retry") else raise Fluent::BigQuery::UnRetryableError.new("failed to insert into bigquery(insert errors), and cannot retry") end end end rescue Google::Apis::ServerError, Google::Apis::ClientError, Google::Apis::AuthorizationError => e @client = nil reason = e.respond_to?(:reason) ? e.reason : nil error_data = { project_id: project, dataset: dataset, table: table_id, code: e.status_code, message: e., reason: reason } wrapped = Fluent::BigQuery::Error.wrap(e) if wrapped.retryable? log.warn "tabledata.insertAll API", error_data else log.error "tabledata.insertAll API", error_data end raise wrapped end |
#wait_load_job(chunk_id, project, dataset, job_id, table_id) ⇒ Object
198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 |
# File 'lib/fluent/plugin/bigquery/writer.rb', line 198 def wait_load_job(chunk_id, project, dataset, job_id, table_id) wait_interval = 10 _response = client.get_job(project, job_id) until _response.status.state == "DONE" log.debug "wait for load job finish", state: _response.status.state, job_id: _response.job_reference.job_id sleep wait_interval _response = client.get_job(project, _response.job_reference.job_id) end errors = _response.status.errors if errors errors.each do |e| log.error "job.insert API (rows)", job_id: job_id, project_id: project, dataset: dataset, table: table_id, message: e., reason: e.reason end end error_result = _response.status.error_result if error_result log.error "job.insert API (result)", job_id: job_id, project_id: project, dataset: dataset, table: table_id, message: error_result., reason: error_result.reason if Fluent::BigQuery::Error.retryable_error_reason?(error_result.reason) @num_errors_per_chunk[chunk_id] = @num_errors_per_chunk[chunk_id].to_i + 1 raise Fluent::BigQuery::RetryableError.new("failed to load into bigquery, retry") else @num_errors_per_chunk.delete(chunk_id) raise Fluent::BigQuery::UnRetryableError.new("failed to load into bigquery, and cannot retry") end end log.debug "finish load job", state: _response.status.state end |