mirror of
https://github.com/octoleo/restic.git
synced 2024-11-30 08:44:02 +00:00
173 lines
5.7 KiB
Go
173 lines
5.7 KiB
Go
|
// Copyright 2016 Google Inc. All Rights Reserved.
|
||
|
//
|
||
|
// Licensed under the Apache License, Version 2.0 (the "License");
|
||
|
// you may not use this file except in compliance with the License.
|
||
|
// You may obtain a copy of the License at
|
||
|
//
|
||
|
// http://www.apache.org/licenses/LICENSE-2.0
|
||
|
//
|
||
|
// Unless required by applicable law or agreed to in writing, software
|
||
|
// distributed under the License is distributed on an "AS IS" BASIS,
|
||
|
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
|
||
|
// See the License for the specific language governing permissions and
|
||
|
// limitations under the License.
|
||
|
|
||
|
package bigquery
|
||
|
|
||
|
import (
|
||
|
"io"
|
||
|
|
||
|
bq "google.golang.org/api/bigquery/v2"
|
||
|
)
|
||
|
|
||
|
// A ReaderSource is a source for a load operation that gets
|
||
|
// data from an io.Reader.
|
||
|
type ReaderSource struct {
|
||
|
r io.Reader
|
||
|
FileConfig
|
||
|
}
|
||
|
|
||
|
// NewReaderSource creates a ReaderSource from an io.Reader. You may
|
||
|
// optionally configure properties on the ReaderSource that describe the
|
||
|
// data being read, before passing it to Table.LoaderFrom.
|
||
|
func NewReaderSource(r io.Reader) *ReaderSource {
|
||
|
return &ReaderSource{r: r}
|
||
|
}
|
||
|
|
||
|
func (r *ReaderSource) populateInsertJobConfForLoad(conf *insertJobConf) {
|
||
|
conf.media = r.r
|
||
|
r.FileConfig.populateLoadConfig(conf.job.Configuration.Load)
|
||
|
}
|
||
|
|
||
|
// FileConfig contains configuration options that pertain to files, typically
|
||
|
// text files that require interpretation to be used as a BigQuery table. A
|
||
|
// file may live in Google Cloud Storage (see GCSReference), or it may be
|
||
|
// loaded into a table via the Table.LoaderFromReader.
|
||
|
type FileConfig struct {
|
||
|
// SourceFormat is the format of the GCS data to be read.
|
||
|
// Allowed values are: CSV, Avro, JSON, DatastoreBackup. The default is CSV.
|
||
|
SourceFormat DataFormat
|
||
|
|
||
|
// FieldDelimiter is the separator for fields in a CSV file, used when
|
||
|
// reading or exporting data. The default is ",".
|
||
|
FieldDelimiter string
|
||
|
|
||
|
// The number of rows at the top of a CSV file that BigQuery will skip when
|
||
|
// reading data.
|
||
|
SkipLeadingRows int64
|
||
|
|
||
|
// AllowJaggedRows causes missing trailing optional columns to be tolerated
|
||
|
// when reading CSV data. Missing values are treated as nulls.
|
||
|
AllowJaggedRows bool
|
||
|
|
||
|
// AllowQuotedNewlines sets whether quoted data sections containing
|
||
|
// newlines are allowed when reading CSV data.
|
||
|
AllowQuotedNewlines bool
|
||
|
|
||
|
// Indicates if we should automatically infer the options and
|
||
|
// schema for CSV and JSON sources.
|
||
|
AutoDetect bool
|
||
|
|
||
|
// Encoding is the character encoding of data to be read.
|
||
|
Encoding Encoding
|
||
|
|
||
|
// MaxBadRecords is the maximum number of bad records that will be ignored
|
||
|
// when reading data.
|
||
|
MaxBadRecords int64
|
||
|
|
||
|
// IgnoreUnknownValues causes values not matching the schema to be
|
||
|
// tolerated. Unknown values are ignored. For CSV this ignores extra values
|
||
|
// at the end of a line. For JSON this ignores named values that do not
|
||
|
// match any column name. If this field is not set, records containing
|
||
|
// unknown values are treated as bad records. The MaxBadRecords field can
|
||
|
// be used to customize how bad records are handled.
|
||
|
IgnoreUnknownValues bool
|
||
|
|
||
|
// Schema describes the data. It is required when reading CSV or JSON data,
|
||
|
// unless the data is being loaded into a table that already exists.
|
||
|
Schema Schema
|
||
|
|
||
|
// Quote is the value used to quote data sections in a CSV file. The
|
||
|
// default quotation character is the double quote ("), which is used if
|
||
|
// both Quote and ForceZeroQuote are unset.
|
||
|
// To specify that no character should be interpreted as a quotation
|
||
|
// character, set ForceZeroQuote to true.
|
||
|
// Only used when reading data.
|
||
|
Quote string
|
||
|
ForceZeroQuote bool
|
||
|
}
|
||
|
|
||
|
// quote returns the CSV quote character, or nil if unset.
|
||
|
func (fc *FileConfig) quote() *string {
|
||
|
if fc.ForceZeroQuote {
|
||
|
quote := ""
|
||
|
return "e
|
||
|
}
|
||
|
if fc.Quote == "" {
|
||
|
return nil
|
||
|
}
|
||
|
return &fc.Quote
|
||
|
}
|
||
|
|
||
|
func (fc *FileConfig) populateLoadConfig(conf *bq.JobConfigurationLoad) {
|
||
|
conf.SkipLeadingRows = fc.SkipLeadingRows
|
||
|
conf.SourceFormat = string(fc.SourceFormat)
|
||
|
conf.Autodetect = fc.AutoDetect
|
||
|
conf.AllowJaggedRows = fc.AllowJaggedRows
|
||
|
conf.AllowQuotedNewlines = fc.AllowQuotedNewlines
|
||
|
conf.Encoding = string(fc.Encoding)
|
||
|
conf.FieldDelimiter = fc.FieldDelimiter
|
||
|
conf.IgnoreUnknownValues = fc.IgnoreUnknownValues
|
||
|
conf.MaxBadRecords = fc.MaxBadRecords
|
||
|
if fc.Schema != nil {
|
||
|
conf.Schema = fc.Schema.asTableSchema()
|
||
|
}
|
||
|
conf.Quote = fc.quote()
|
||
|
}
|
||
|
|
||
|
func (fc *FileConfig) populateExternalDataConfig(conf *bq.ExternalDataConfiguration) {
|
||
|
format := fc.SourceFormat
|
||
|
if format == "" {
|
||
|
// Format must be explicitly set for external data sources.
|
||
|
format = CSV
|
||
|
}
|
||
|
// TODO(jba): support AutoDetect.
|
||
|
conf.IgnoreUnknownValues = fc.IgnoreUnknownValues
|
||
|
conf.MaxBadRecords = fc.MaxBadRecords
|
||
|
conf.SourceFormat = string(format)
|
||
|
if fc.Schema != nil {
|
||
|
conf.Schema = fc.Schema.asTableSchema()
|
||
|
}
|
||
|
if format == CSV {
|
||
|
conf.CsvOptions = &bq.CsvOptions{
|
||
|
AllowJaggedRows: fc.AllowJaggedRows,
|
||
|
AllowQuotedNewlines: fc.AllowQuotedNewlines,
|
||
|
Encoding: string(fc.Encoding),
|
||
|
FieldDelimiter: fc.FieldDelimiter,
|
||
|
SkipLeadingRows: fc.SkipLeadingRows,
|
||
|
Quote: fc.quote(),
|
||
|
}
|
||
|
}
|
||
|
}
|
||
|
|
||
|
// DataFormat describes the format of BigQuery table data.
|
||
|
type DataFormat string
|
||
|
|
||
|
// Constants describing the format of BigQuery table data.
|
||
|
const (
|
||
|
CSV DataFormat = "CSV"
|
||
|
Avro DataFormat = "AVRO"
|
||
|
JSON DataFormat = "NEWLINE_DELIMITED_JSON"
|
||
|
DatastoreBackup DataFormat = "DATASTORE_BACKUP"
|
||
|
)
|
||
|
|
||
|
// Encoding specifies the character encoding of data to be loaded into BigQuery.
|
||
|
// See https://cloud.google.com/bigquery/docs/reference/v2/jobs#configuration.load.encoding
|
||
|
// for more details about how this is used.
|
||
|
type Encoding string
|
||
|
|
||
|
const (
|
||
|
UTF_8 Encoding = "UTF-8"
|
||
|
ISO_8859_1 Encoding = "ISO-8859-1"
|
||
|
)
|