PostgreSQL Dynamic SQL – How to Create a Table Based on JSON Data

dynamic-sqlfunctionsjsonpostgresql

I have a json that I get as a file, I have loaded that in the Postgres database.
Here is the json data:

{
"tablename": "test",
"columns": [
    {
        "name": "field1",
        "datatype": "BigInt"
    },
    {
        "name": "field2",
        "datatype": "String"
    }
]

}

Now I have to create a table dynamically, I am thinking of writing a function in Postgres to do that.
So the table would be named test with 2 fields one as string and another as bigint.

I am able to get the table name by doing a select as below:

select (metadata->'tablename') from public.json_metadata;

However, I am having difficulty getting all the nested column names to form a create table statement.

1- How would you go about doing that, any built in Postgres functions to extract that.

2- Is a Postgres function the best way to approach this problem, or should I write this in python (I will have to learn Python) or shell script.

The number of columns would not be fixed, different json files will have different number of columns.

Best Answer

You need dynamic SQL for that, which carries the hazard of SQL injection.
However, done properly, this is safe against SQLi:

DO
$do$
BEGIN

EXECUTE (
   SELECT format('CREATE TABLE %I(%s)', metadata->>'tablename', c.cols)
   FROM   public.json_metadata m
   CROSS  JOIN LATERAL (
      SELECT string_agg(quote_ident(col->>'name')
                        || ' ' ||  (col->>'datatype')::regtype, ', ') AS cols
      FROM   json_array_elements(metadata->'columns') col
      ) c
   );
END
$do$;

Table and column names are treated as case-sensitive. (You may want lower-case instead.) The type name is treated as case-insensitive, and any valid type name works.

Of course, this would raise an exception for the non-existent data type String you display. Try with text instead.

Note the use of format(), the object identifier type regclass, quote_ident(), the aggregation of columns in the LATERAL subquery and the DO command to execute dynamic SQL.

PostgreSQL – How to Turn JSON Array into Postgres Array?

Postgres 9.4 or newer

Obviously inspired by this post, Postgres 9.4 added the missing function(s):
_{Thanks to Laurence Rowe for the patch and Andrew Dunstan for committing!}

To unnest the JSON array. Then use array_agg() or an ARRAY constructor to build a Postgres array from it. Or string_agg() to build a text string.

Aggregate unnested elements per row in a LATERAL or correlated subquery. Then original order is preserved and we don't need ORDER BY, GROUP BY or even a unique key in the outer query. See:

How to apply ORDER BY and LIMIT in combination with an aggregate function?

Replace 'json' with 'jsonb' for jsonb in all following SQL code.

SELECT t.tbl_id, d.list
FROM   tbl t
CROSS  JOIN LATERAL (
   SELECT string_agg(d.elem::text, ', ') AS list
   FROM   json_array_elements_text(t.data->'tags') AS d(elem)
   ) d;

Short syntax:

SELECT t.tbl_id, d.list
FROM   tbl t, LATERAL (
   SELECT string_agg(value::text, ', ') AS list
   FROM   json_array_elements_text(t.data->'tags')  -- col name default: "value"
   ) d;

What is the difference between LATERAL and a subquery in PostgreSQL?

ARRAY constructor in correlated subquery:

SELECT tbl_id, ARRAY(SELECT json_array_elements_text(t.data->'tags')) AS txt_arr
FROM   tbl t;

How to apply ORDER BY and LIMIT in combination with an aggregate function?

Subtle difference: null elements are preserved in actual arrays. This is not possible in the above queries producing a text string, which cannot contain null values. The true representation is an array.

Function wrapper

For repeated use, to make this even simpler, encapsulate the logic in a function:

CREATE OR REPLACE FUNCTION json_arr2text_arr(_js json)
  RETURNS text[] LANGUAGE sql IMMUTABLE PARALLEL SAFE AS
'SELECT ARRAY(SELECT json_array_elements_text(_js))';

Make it an SQL function, so it can be inlined in bigger queries.
Make it IMMUTABLE (because it is) to avoid repeated evaluation in bigger queries and allow it in index expressions.
Make it PARALLEL SAFE (in Postgres 9.6 or later!) to not stand in the way of parallelism. See:

When to mark functions as PARALLEL RESTRICTED vs PARALLEL SAFE?

Call:

SELECT tbl_id, json_arr2text_arr(data->'tags')
FROM   tbl;

db<>fiddle here

Postgres 9.3 or older

Use the function json_array_elements(). But we get double quoted strings from it.

Alternative query with aggregation in the outer query. CROSS JOIN removes rows with missing or empty arrays. May also be useful for processing elements. We need a unique key to aggregate:

SELECT t.tbl_id, string_agg(d.elem::text, ', ') AS list
FROM   tbl t
CROSS  JOIN LATERAL json_array_elements(t.data->'tags') AS d(elem)
GROUP  BY t.tbl_id;

ARRAY constructor, still with quoted strings:

SELECT tbl_id, ARRAY(SELECT json_array_elements(t.data->'tags')) AS quoted_txt_arr
FROM   tbl t;

Note that null is converted to the text value "null", unlike above. Incorrect, strictly speaking, and potentially ambiguous.

Poor man's unquoting with trim():

SELECT t.tbl_id, string_agg(trim(d.elem::text, '"'), ', ') AS list
FROM   tbl t, json_array_elements(t.data->'tags') d(elem)
GROUP  BY 1;

Retrieve a single row from tbl:

SELECT string_agg(trim(d.elem::text, '"'), ', ') AS list
FROM   tbl t, json_array_elements(t.data->'tags') d(elem)
WHERE  t.tbl_id = 1;

Strings form correlated subquery:

SELECT tbl_id, (SELECT string_agg(trim(value::text, '"'), ', ')
                FROM   json_array_elements(t.data->'tags')) AS list
FROM   tbl t;

ARRAY constructor:

SELECT tbl_id, ARRAY(SELECT trim(value::text, '"')
                     FROM   json_array_elements(t.data->'tags')) AS txt_arr
FROM   tbl t;

Original (outdated) SQL Fiddle.
db<>fiddle here.

Need to select a JSON array element dynamically from a postgresql table

Notes (outdated since pg 9.4)

We would need a json_array_elements_text(json), the twin of json_array_elements(json) to return proper text values from a JSON array. But that seems to be missing from the provided arsenal of JSON functions. Or some other function to extract a text value from a scalar JSON value. I seem to be missing that one, too.
So I improvised with trim(), but that will fail for non-trivial cases ...

PostgreSQL Trigger Function: Selecting from Current Table

I suggested that you use trigger arguments, but it's actually not necessary. You can use the automatic variables TG_TABLE_SCHEMA and TG_TABLE_NAME, or use TG_RELID. These, alongside EXECUTE for dynamic SQL, let you do what you want:

BEGIN
    EXECUTE format('SELECT colname FROM %I', TG_RELID)
END;

BEGIN
    EXECUTE format('SELECT colname FROM %I.%I', TG_TABLE_SCHEMA, TG_TABLE_NAME)
END;

(Of course these won't work as-is, since the SELECT has no destination for the data. You have to use EXECUTE format(..) INTO ... to store the result into a DECLAREd variable), e.g.

DECLARE
    _colvar integer;
BEGIN
    EXECUTE format('SELECT colname FROM %I.%I', TG_TABLE_SCHEMA, TG_TABLE_NAME) INTO _colvar;
    RAISE NOTICE 'colname value was %',_colvar;
END;