CODEWITHSUNDEEP
jsonlualua-tabletarantool
Anh Duy Tran Do
Anh Duy Tran Do

Reputation: 33

How to insert a LUA Table to Tarantool with Key and Value?

I have a JSON string:

{
"entry_offset" : 180587225765,
"entry_size" : 54003,
"created_time" : 1577500878,
"additional_meta" : {
    "geohash64" : 5637765837143565,
    "mime_type" : "image/jpg"
}

And I have converted it to Lua Table using Tarantool's module json:

table = json.decode(JSONstring)

Then I want to insert the table to Tarantool with ID = 1

box.space.somespace:insert{1, table}

The result is like this when I select the table which added to Tarantool database in JSON form:

Cannot access values through key

I can only access to table[1] and table[2]: table[1] is the ID = 1 while table[2] is all the JSON string. Which means I can't access the Values of JSON with Keys: table['entry_offset'], table['entry_size'], .... return nil when I try to access them

So how can I insert a Lua table to Tarantool and then access the values through its keys?

Thanh you so much for helping!!!

Upvotes: 1

Views: 690

Answers (1)

RunFor
RunFor

Reputation: 555

You are basically inserting a wrapping tuple into the space, not table object itself, so when you do:

obj = box.space.somespace:get{1}

you get your tuple back, not table. That said if you want to access table's fields using keys, you just need to index that obj like this:

table = obj[2]
print(table.entry_offset)

Once you get used to it, check out tarantool space format feature with its tomap/frommap functions. Here is basic example, which may help you:

box.cfg{}
box.schema.create_space('test', {if_not_exists = true})
box.space.test:create_index('pk', {unique = true, if_not_exists = true, parts = {1, 'unsigned'}})
box.space.test:format({
    { name = 'id', 'unsigned' },
    { name = 'entry_offset', 'unsigned' },
    { name = 'entry_size', 'unsigned' },
    { name = 'created_time', 'unsigned' },
    { name = 'additional_meta', 'map' },
})

json = require('json')
obj_json = [[{
"entry_offset" : 180587225765,
"entry_size" : 54003,
"created_time" : 1577500878,
"additional_meta" : {
    "geohash64" : 5637765837143565,
    "mime_type" : "image/jpg"
}}]]
obj = json.decode(obj_json)
obj.id = 1

tuple = box.space.test:frommap(obj)
box.space.test:insert(tuple)
result = box.space.test:get({1}):tomap()

print(result.additional_meta.mime_type)

Or, for more advanced serialization approach, take a look at avro-schema with its flatten/unflatten methods. There are examples in README

Upvotes: 2

Related Questions