avro

The purpose of this package is to facilitate use of AVRO with go strong typing.

Features

github.com/khezen/avro

• Marshal/Unmarshal AVRO schema

github.com/khezen/avro/sqlavro

• Discover SQL tables
• Convert SQL tables to AVRO schemas
• Query records from SQL into AVRO or CSV binary

github.com/khezen/avro/redshiftavro

• Produce Redshit create statement from AVRO schema

What is AVRO

Apache AVRO is a data serialization system which relies on JSON schemas.

It provides:

• Rich data structures
• A compact, fast, binary data format
• A container file, to store persistent data
• Remote procedure call (RPC)

AVRO binary encoded data comes together with its schema and therefore is fully self-describing.

When AVRO data is read, the schema used when writing it is always present. This permits each datum to be written with no per-value overheads, making serialization both fast and small.

When AVRO data is stored in a file, its schema is stored with it, so that files may be processed later by any program. If the program reading the data expects a different schema this can be easily resolved, since both schemas are present.

Examples

Schema Marshal/Unmarshal

package main

import (
  "encoding/json"
  "fmt"

  "github.com/khezen/avro"
)

func main() {
  schemaBytes := []byte(
    `{
      "type": "record",
      "namespace": "test",
      "name": "LongList",
      "aliases": [
        "LinkedLongs"
      ],
      "doc": "linked list of 64 bits integers",
      "fields": [
        {
          "name": "value",
          "type": "long"
        },
        {
          "name": "next",
          "type": [
            "null",
            "LongList"
          ]
        }
      ]
    }`,
  )

  // Unmarshal JSON  bytes to Schema interface
  var anySchema avro.AnySchema
  err := json.Unmarshal(schemaBytes, &anySchema)
  if err != nil {
    panic(err)
  }
  schema := anySchema.Schema()  
  // Marshal Schema interface to JSON bytes
  schemaBytes, err = json.Marshal(schema)
  if err != nil {
    panic(err)
  }
  fmt.Println(string(schemaBytes))
}

{
    "type": "record",
    "namespace": "test",
    "name": "LongList",
    "aliases": [
        "LinkedLongs"
    ],
    "doc": "linked list of 64 bits integers",
    "fields": [
        {
            "name": "value",
            "type": "long"
        },
        {
            "name": "next",
            "type": [
                "null",
                "LongList"
            ]
        }
    ]
}

Convert SQL Table to AVRO Schema

package main
import (
  "database/sql"
  "encoding/json"
  "fmt"

  "github.com/khezen/avro/sqlavro"
)

func main() {
  db, err := sql.Open("mysql", "root@/blog")
  if err != nil {
    panic(err)
  }
  defer db.Close()
  _, err = db.Exec(
    `CREATE TABLE posts(
      ID INT NOT NULL,
      title VARCHAR(128) NOT NULL,
      body LONGBLOB NOT NULL,
      content_type VARCHAR(128) DEFAULT 'text/markdown; charset=UTF-8',
      post_date DATETIME NOT NULL,
      update_date DATETIME,
      reading_time_minutes DECIMAL(3,1),
      PRIMARY KEY(ID)
    )`,
  )
  if err != nil {
    panic(err)
  }
  schemas, err := sqlavro.SQLDatabase2AVRO(db, "blog")
  if err != nil {
    panic(err)
  }
  schemasBytes, err := json.Marshal(schemas)
  if err != nil {
    panic(err)
  }
  fmt.Println(string(schemasBytes))
}

[
    {
        "type": "record",
        "namespace": "blog",
        "name": "posts",
        "fields": [
            {
                "name": "ID",
                "type": "int"
            },
            {
                "name": "title",
                "type": "string"
            },
            {
                "name": "body",
                "type": "bytes"
            },
            {
                "name": "content_type",
                "type": [
                    "string",
                    "null"
                ],
                "default": "text/markdown; charset=UTF-8"
            },
            {
                "name": "post_date",
                "type": {
                    "type": "int",
                    "doc":"datetime",
                    "logicalType": "timestamp"
                }
            },
            {
                "name": "update_date",
                "type": [
                    "null",
                    {
                        "type": "int",
                        "doc":"datetime",
                        "logicalType": "timestamp"
                    }
                ]
            },
            {
                "name": "reading_time_minutes",
                "type": [
                    "null",
                    {
                        "type": "bytes",
                        "logicalType": "decimal",
                        "precision": 3,
                        "scale": 1
                    }
                ]
            }
        ]
    }
]

Query records from SQL into AVRO or CSV binary

package main

import (
	"database/sql"
	"fmt"
	"io/ioutil"
	"time"

	"github.com/khezen/avro"
	"github.com/khezen/avro/sqlavro"
)

func main() {
	db, err := sql.Open("mysql", "root@/blog")
	if err != nil {
		panic(err)
	}
	defer db.Close()
	_, err = db.Exec(
		`CREATE TABLE posts(
			ID INT NOT NULL,
			title VARCHAR(128) NOT NULL,
			body LONGBLOB NOT NULL,
			content_type VARCHAR(128) DEFAULT 'text/markdown; charset=UTF-8',
			post_date DATETIME NOT NULL,
			update_date DATETIME,
			reading_time_minutes DECIMAL(3,1),
			PRIMARY KEY(ID)
		)`,
	)
	if err != nil {
		panic(err)
	}
	_, err = db.Exec(
		// statement
		`INSERT INTO posts(ID,title,body,content_type,post_date,update_date,reading_time_minutes)
		 VALUES (?,?,?,?,?,?,?)`,
		// values
		42,
		"lorem ispum",
		[]byte(`Lorem ipsum dolor sit amet, consectetur adipiscing elit, sed do eiusmod tempor incididunt ut labore et dolore magna aliqua.`),
		"text/markdown; charset=UTF-8",
		"2009-04-10 00:00:00",
		"2009-04-10 00:00:00",
		"4.2",
	)
	if err != nil {
		panic(err)
	}
	schema, err := sqlavro.SQLTable2AVRO(db, "blog", "posts")
	if err != nil {
		panic(err)
	}
	limit := 1000
	order := avro.Ascending
	from, err := time.Parse("2006-02-01 15:04:05", "2009-04-10 00:00:00")
	if err != nil {
		panic(err)
	}
	avroBytes, updatedCriteria, err := sqlavro.Query(sqlavro.QueryConfig{
		DB:     db,
		DBName: "blog",
		Schema: schema,
		Limit:  limit,
		Criteria: []sqlavro.Criterion{
			*sqlavro.NewCriterionDateTime("post_date", &from, order),
		},
		Output: "avro",
	})
	if err != nil {
		panic(err)
	}
	err = ioutil.WriteFile("/tmp/blog_posts.avro", avroBytes, 0644)
	if err != nil {
		panic(err)
	}
	fmt.Println(updatedCriteria)
}

Notes

• When record fields contains aliases, the first alias is used in the query instead of the field name.

Types

Because of encoding rules for Avro unions, when an union's value is null, a simple Go nil is returned. However when an union's value is non-nil, a Go map[string]interface{} with a single key is returned for the union. The map's single key is the Avro type name and its value is the datum's value.

Produce Redshift create statement from AVRO schema

package main

import (
	"encoding/json"
	"fmt"

	"github.com/khezen/avro"
	"github.com/khezen/avro/redshiftavro"
)

func main() {
	schemaBytes := []byte(`
	{
        "type": "record",
        "namespace": "blog",
        "name": "posts",
        "fields": [
            {
                "name": "ID",
                "type": "int"
            },
            {
                "name": "title",
                "type": "string"
            },
            {
                "name": "body",
                "type": "bytes"
            },
            {
                "name": "content_type",
                "type": [
                    "string",
                    "null"
                ],
                "default": "text/markdown; charset=UTF-8"
            },
            {
                "name": "post_date",
                "type": {
                    "type": "int",
                    "doc":"datetime",
                    "logicalType": "timestamp"
                }
            },
            {
                "name": "update_date",
                "type": [
                    "null",
                    {
                        "type": "int",
                        "doc":"datetime",
                        "logicalType": "timestamp"
                    }
                ]
            },
            {
                "name": "reading_time_minutes",
                "type": [
                    "null",
                    {
                        "type": "bytes",
                        "logicalType": "decimal",
                        "precision": 3,
                        "scale": 1
                    }
                ]
            }
        ]
	}`)
	var anySchema avro.AnySchema
	err := json.Unmarshal(schemaBytes, &anySchema)
	if err != nil {
		panic(err)
	}
	schema := anySchema.Schema().(*avro.RecordSchema)
	cfg := redshiftavro.CreateConfig{
		Schema:      *schema,
		SortKeys:    []string{"post_date", "title"},
		IfNotExists: true,
	}
	statement, err := redshiftavro.CreateTableStatement(cfg)
	if err != nil {
		panic(err)
	}
	fmt.Println(statement)
}

CREATE TABLE IF NOT EXISTS posts(
	ID INTEGER ENCODE LZO NOT NULL,
	title VARCHAR(65535) ENCODE RAW NOT NULL,
	body VARCHAR(65535) ENCODE ZSTD NOT NULL,
	content_type VARCHAR(65535) ENCODE ZSTD NULL,
	post_date TIMESTAMP WITHOUT TIME ZONE ENCODE RAW NOT NULL,
	update_date TIMESTAMP WITHOUT TIME ZONE ENCODE LZO NULL,
	reading_time_minutes DECIMAL(3,1) ENCODE RAW NULL
)
SORTKEY(
	post_date,
	title
)

Issues

If you have any problems or questions, please ask for help through a GitHub issue.

Contributions

Help is always welcome! For example, documentation (like the text you are reading now) can always use improvement. There's always code that can be improved. If you ever see something you think should be fixed, you should own it. If you have no idea what to start on, you can browse the issues labeled with help wanted.

As a potential contributor, your changes and ideas are welcome at any hour of the day or night, weekdays, weekends, and holidays. Please do not ever hesitate to ask a question or send a pull request.

Code of conduct.