Prohibit lines from being exactly the same in the database

Question

I have a system that integrates data from three different banks, more specifically the registration of the companies of these three banks, I created a table where are stored the integrations that user will do, for example:

no banco X a empresa A é código 01
no banco Y a empresa A é código 02
no banco Z a empresa A é código 03

in my table that stores the integrations that user will do the data would look like this:

nomeEmpresa = A
codigoBancoX = 1
codigoBancoY = 2
codigoBancoZ = 3

I need to avoid that by carelessness of the user be added exactly equal all fields, ie can not have two rows in my table exactly equal. My question is. If I put all fields of my table as Primary Key would solve my problem? Because then no line could be exactly the same but still could exist line like this:

nomeEmpresa = A
codigoBancoX = 1
codigoBancoY = 2
codigoBancoZ = 3

nomeEmpresa = A
codigoBancoX = 1
codigoBancoY = 2
codigoBancoZ = 4

Answer 1

As stated in the comments, the correct situation is to use a single index.

Create a UNIQUE INDEX in postgresql:

CREATE UNIQUE INDEX tabela_unique_idx ON public.tabela
  USING btree ("col1", "col2", "col3");

Reference: https://www.postgresql.org/docs/9.4/static/indexes-unique.html

Answer 2

PK uniquely identifies each record in a database table, UNIQUE recommended for what you need. and you have to make error treatment in your application when the records are identical.

ERROR 1062: 1062: Duplicate entry 'value' for key 'column_UNIQUE'

Answer 3

From your description, I believe you need each field to have a unique value, after all you shouldn’t have two companies in the banco X with the same id.

To do this, you must create a unique Constraint for each field.:

ALTER TABLE public.Empresas ADD CONSTRAINT CK_Empresas_codigoBancoX UNIQUE (codigoBancoX);
ALTER TABLE public.Empresas ADD CONSTRAINT CK_Empresas_codigoBancoY UNIQUE (codigoBancoY);
ALTER TABLE public.Empresas ADD CONSTRAINT CK_Empresas_codigoBancoZ UNIQUE (codigoBancoZ);

your second option is to use a unique Intel

CREATE UNIQUE INDEX IXU_Empresas_codigoBancoX ON public.Empresas (codigoBancoX);
CREATE UNIQUE INDEX IXU_Empresas_codigoBancoX ON public.Empresas (codigoBancoY);
CREATE UNIQUE INDEX IXU_Empresas_codigoBancoX ON public.Empresas (codigoBancoZ);

would say to create an Index if this code is used as a condition in any query.