BUG #15946: "duplicate key" error on ANALYZE of table partitions in transaction

The following bug has been logged on the website:

Bug reference:      15946
Logged by:          Yaroslav Schekin
Email address:      ladayaroslav@yandex.ru
PostgreSQL version: 11.5
Operating system:   Any
Description:

Running this:
-----
BEGIN TRANSACTION;
CREATE TABLE sg (
id bigint NOT NULL,
sc_fk bigint,
geo_id bigint,
sl smallint NOT NULL,
a date NOT NULL,
o boolean NOT NULL
)
PARTITION BY RANGE (o, sl, a);

CREATE TABLE sg_19_01_d PARTITION OF sg FOR VALUES FROM (false, '5',
'2019-01-01') TO (false, '5', '2019-02-01');
CREATE TABLE sg_19_02_d PARTITION OF sg FOR VALUES FROM (false, '5',
'2019-02-01') TO (false, '5', '2019-03-01');

CREATE TABLE sc (
id bigint,
a date NOT NULL,
sl smallint NOT NULL,
o boolean NOT NULL
)
PARTITION BY RANGE (o, sl, a);

CREATE TABLE sc_19_01_d PARTITION OF sc FOR VALUES FROM (false, '5',
'2019-01-01') TO (false, '5', '2019-02-01');
CREATE TABLE sc_19_02_d PARTITION OF sc FOR VALUES FROM (false, '5',
'2019-02-01') TO (false, '5', '2019-03-01');

INSERT INTO sg_19_01_d(id, sc_fk, geo_id, sl, a, o)
SELECT n, n, 0, 5, '2019-01-01', false
  FROM generate_series(1, 1000) AS g(n);

INSERT INTO sg_19_02_d(id, sc_fk, geo_id, sl, a, o)
SELECT n, n, 0, 5, '2019-02-01', false
  FROM generate_series(1, 1000) AS g(n);

INSERT INTO sc_19_01_d(id, a, sl, o)
SELECT n, '2019-01-01', 5, false
  FROM generate_series(1, 1000) AS g(n);

INSERT INTO sc_19_02_d(id, a, sl, o)
SELECT n, '2019-02-01', 5, false
  FROM generate_series(1, 1000) AS g(n);

ANALYZE sg_19_01_d, sc_19_02_d, sc_19_01_d, sc_19_02_d;
-----
Throws this error:
ERROR:  duplicate key value violates unique constraint
"pg_statistic_relid_att_inh_index"
DETAIL:  Key (starelid, staattnum, stainherit)=(61056, 1, f) already exists.