When using the CSV import feature to update existing records in a table, PostgreSQL throws a “This record already exists” error. This error does not occur when using MySQL.
Example:
test table，title is PRIMARY KEY，the new data is
title,name
1,5
2,5
3,3
PG ‘s CREATE TABLE SQL:
CREATE TABLE test (
“id” int4 NOT NULL DEFAULT nextval(‘test_id_seq’::regclass),
“created_at” timestamp(6),
“updated_at” timestamp(6),
“created_by” varchar COLLATE “pg_catalog”.“default”,
“updated_by” varchar COLLATE “pg_catalog”.“default”,
“nc_order” numeric,
“title” text COLLATE “pg_catalog”.“default” NOT NULL,
“name” int8,
CONSTRAINT “test_pkey” PRIMARY KEY (“title”)
)
this is caused by differences in how MySQL and PostgreSQL handle upsert operations (updating existing records based on a primary key). Switching from using the bulkDataInsert API to the bulkDataUpsert API should resolve this problem.
I am not familiar with the NocoDB codebase and am unsure where to make this specific code change.
