Postgresql Many to Many relationship insert

Question

I'm using Postresql 9.3.5. I have a many-to-many relationship between entities "Foo" and "Bar" that I've modeled as something like:

CREATE TABLE Foo
(
    id SERIAL PRIMARY KEY NOT NULL,
    .... various columns for foo ....
);

CREATE TABLE Bar
(
    id SERIAL PRIMARY KEY NOT NULL,
    field1 varchar(50) UNIQUE NOT NULL,
    .... various columns for bar ....
);

CREATE TABLE FooBar
(
    fooID int NOT NULL,
    barID int NOT NULL,
    PRIMARY KEY (fooID, barID),
    FOREIGN KEY (fooID) REFERENCES Foo(id),
    FOREIGN KEY (barID) REFERENCES Bar(id)
);

Now what I want to do is insert a record into Foo, insert a corresponding record into Bar, and then insert the matching FooBar record containing the ids of the foo & bar entries.

Catch: I don't know when I go to insert the Bar records if they already exist, so currently my insert for Bar looks something like:

INSERT INTO Bar(field1, .... other fields for Bar....) 
SELECT 'value1', .... other values for the insert....
WHERE NOT EXISTS (
SELECT 1 FROM Bar WHERE field1 = 'value1')

Which works fine, but my question: how do I get the id of the newly inserted (or existing) Bar record so that I can insert it into the FooBar table?

Answer 1

This seems to work, although it is far from elegant and is probably very inefficient:

WITH new AS (
    INSERT INTO bar(field1) 
    SELECT ('aaa') WHERE NOT EXISTS (
        SELECT 1 FROM bar WHERE field1='aaa'
    ) RETURNING id
),
existing AS (
    SELECT id FROM bar WHERE field1='aaa'
) 
SELECT id FROM existing UNION SELECT id FROM new

I imagine this would be inefficient due to repeated searches in bar for the matching value. A more efficient solution might be to write a stored procedure.