SQL question on natural joins
I am a newbie to SQL and I studied the book of Fundamentals SQL Server Oracle OCA 12 c, but a section on natural joins to confuse me.
Basically, the author says:
SELECT r.region_name, d.department_name, l.city, c.country_name
DEPARTMENTS d
NATURAL JOIN places l, c countries, regions r;
The natural join between DEPARTMENTS and LOCATIONS creates a provisional result, consisting of 27 lines since they are implicitly attached on the column of location_id. This set is then Cartesian-joined to the table of the COUNTRY as a join condition is not implicitly or explicitly specified. 27 interim lines are attached to 25 lines in the COUNTRY table, which gives a new temp results set with 675 (27 × 25) rows and three columns: DEPARTMENT_NAME, CITY and COUNTRY_NAME. This set is then attached to the REGION table. Once more, a Cartesian join occurs because column REGION_ID is absent from any join condition. The final result set contains rows and four columns (675 × 4) 2700.
I can understand because you evaluate joins from the left. But then he wrote:
The JOIN... USE and JOIN... ON the syntaxes are better suited to join multiple tables. The following query joins four tables using the natural join syntax:
SELECT country_id region_id, c.country_name, l.city, d.department_name
DEPARTMENTS d
NATURAL JOIN places l
NATURAL JOIN country c
NATURAL JOIN region r;
This query generates correctly 27 lines in the final results set since the required join columns are listed in the SELECT clause. The following query illustrates how the JOIN... CLAUSE is used to extract the same 27 lines. A join condition can reference columns only in its scope. In the following example, the join of DEPARTMENTS slots can not reference columns in the COUNTRIES or REGIONS of tables, but the join between the COUNTRIES and REGIONS may refer to any column of four tables involved in the query.
This second method of the part of the natural writing joined confuses me. I don't understand the logic behind the 2nd series of States of Natural Join. For me, it seems that the first series and the 2nd set look alike, but they are apparently not.
Can someone tell me the differences?
Thank you!!
Hello
The more important thing to learn more about NATURAL JOIN is: never use it. If you add new columns, joins can get different results. I've never heard of someone uisng NATURAL JOIN apart from a manual or a question like yours.
There are a lot of things in Oracle that take the time to learn and are useful. All the time you spend learning things is better spent on them.
Tags: Database
Similar Questions
-
Natural join of Table to herself out of Cartesian product
Hello
With
CREATE TABLE A)
an INTEGER,
b VARCHAR (15).
c DATE
);
INSERT IN A (a, b, c) VALUES (1, 'AAA', SYSDATE);
INSERT IN A (a, b, c) VALUES (2, 'BBBB', SYSDATE);
INSERT IN A (a, b, c) VALUES (3, 'CCCCC', SYSDATE);
CREATE VIEW A1
AS SELECT * FROM A
;
the query
SELECT
*
A.
NATURAL JOIN HAS
;
returning nine records - but
SELECT
*
A.
NATURAL JOIN A1
;
only three - which is what I expected.
Why the difference?
Tested in 11g. Where is the documentation pertaining to this particular aspect?
Thank you!
Report it as a bug, the correct result, it's the three ranks of result you can go back to giving has an alias, for example:
SQL > select * a natural join a and b;
A B C
---------- --------------- ---------
1 AAA 15 JULY 15
2 BBBB 15 JULY 15
3 CARTER 15 JULY 15
3 selected lines.
SQL > select * from a natural join.
A B C
---------- --------------- ---------
1 AAA 15 JULY 15
1 AAA 15 JULY 15
1 AAA 15 JULY 15
2 BBBB 15 JULY 15
2 BBBB 15 JULY 15
2 BBBB 15 JULY 15
3 CARTER 15 JULY 15
3 CARTER 15 JULY 15
3 CARTER 15 JULY 15
9 selected lines.
Run on 11.2.0.4 - but reproduced on 12.1.0.2
Concerning
Jonathan Lewis
-
Incorrect results of natural join producing. I was wondering why.
If anyones got a second, I have a small question about this bit of code. It seems that he should produce good results, but does not work.
"Write a query to identify the highest salary paid in each country. It will take by using a subquery in the FROM clause. »
It is the obvious answer:
"select max (salary), country_id of."
(select department_id, location_id, salary, country_id of)
employees join natural services locations of natural join)
Country_id group; »
Here are the results:
17000 WE
CA 6000
10000 UK
I wrote this instead:
Select max (salary), country_id of
(select d.department_id, l.location_id, c.country_id, e.salary, c.country_name
e employees
Join departments d on (e.department_id = d.department_id)
Join places l on (d.location_id = l.location_id)
Join country c on (l.country_id = c.country_id)
)
Country_id group;
Which produces:
24000 WE
13000 CA
10000 OF
14000 UK
(Of course it also works with 'join using (column_name)')
My answers look correct when passing through the countries/places/employee tables manually - look of the wrong example.
I guess its something to do with the natural join? Maybe his join tables somewhere on several columns? I would like to understand exactly why the first example is if bad - because I sure hope will teach me more about why avoid natural joins.
Also, are there ways to improve outcomes, for example if I was watching an average salary? I drew the tables to try to find ways that the records may be left out, and where the outer joins might be useful. The only thing I could think of would be that a Department/location not assigned a place/country respectively. But I'm not an outer join would allow it at all except if I created some NVL2 code to assign new place/country IDs based on other areas - I think that's a bit much for this example.
Thank you very much
Nick
Published by: nick woodward on April 22, 2013 11:10
Published by: nick woodward on April 22, 2013 11:12Hi, Nick.
Nick woodward wrote:
If anyones got a second, I have a small question about this bit of code. It seems that he should produce good results, but does not work."Write a query to identify the highest salary paid in each country. It will take by using a subquery in the FROM clause. »
In fact, a subquery in the FROM clause is not yet useful, at least not for me, and it is certainly not necessary.
It is the obvious answer:
"select max (salary), country_id of."
(select department_id, location_id, salary, country_id of)
employees join natural services locations of natural join)
Country_id group; »Here are the results:
17000 WE
CA 6000
10000 UKI wrote this instead:
Select max (salary), country_id of
(select d.department_id, l.location_id, c.country_id, e.salary, c.country_name
e employees
Join departments d on (e.department_id = d.department_id)
Join places l on (d.location_id = l.location_id)
Join country c on (l.country_id = c.country_id)
)
Country_id group;Which produces:
24000 WE
13000 CA
10000 OF
14000 UK(Of course it also works with 'join using (column_name)')
AID is not as bad as NATURAL JOIN, but it's pretty bad. Forget the JOIN... Help me, just like you should forget NATURAL JOIN.
My answers look correct when passing through the countries/places/employee tables manually - look of the wrong example.
I guess its something to do with the natural join? Maybe his join tables somewhere on several columns?
Exactly; There is a column manager_id employees and also in the departments.
If the obligation is to get the highest salary of all employees in a country, then NATURAL JOIN does not meet the requirements.I would like to understand exactly why the first example is if bad - because I sure hope will teach me more about why avoid natural joins.
Do you need any other reasons?
Also, are there ways to improve the results
As I said earlier, you don't need a subquery for this.
In addition, you must the country table if you want to display the country_name, or if you need to reach the region table. In this problem, all you need is the country_id and you can get that from the communities table, then the following text also gets good results:SELECT MAX (e.salary) AS max_salary , l.country_id FROM employees e JOIN departments d ON d.department_id = e.department_id JOIN locations l ON l.location_id = d.location_id GROUP BY l.country_id ORDER BY l.country_id ;Whenever a request is for 2 or more tables, it is recommended to describe each column with a table name or alias.
say if I watched an average salary?
You can use AVG much lke you use MAX. You can do both in the same query, if you wish; See the example below.
I drew the tables to try to find ways that the records may be left out, and where the outer joins might be useful. The only thing I could think of would be that a Department/location not assigned a place/country respectively.
Right; outer joins are useful when you want to get data from the table or not a matching rows in the table b. Departaments that have not been allocated a location is a situation that calls for an outer join. Another example is the country who have no departments in them. (There is a real possibility. You can set up a table with all countries in the world already in it, not because you'll need all of them, but because you might need one of them.)
But I'm not an outer join would allow it at all except if I created some NVL2 code to assign new place/country IDs based on other areas - I think that's a bit much for this example.
Not necessarily. Sometimes all you need is the NULL that you get automatically for all columns of table b when you say
FROM a LEFT OUTER JOIN b ...and one has no corresponding row in the b
For example, if you want to include all countries in the table of localities, with the salary maximum and average of those that have employees, you can get these results:
MAX_SALARY AVG_SALARY CO ---------- ---------- -- AU BR 13000 9500 CA CH CN 10000 10000 DE IN IT JP MX NL SG 14000 8885.71429 UK 24000 5064.70588 USwithout using NVL, or NVL2 or something like that. In fact, only the functions you need are AVG and MAX.
Try it.Published by: Frank Kulash on April 22, 2013 16:35
-
the two equii join and natural join are equall.will both display the output of the same
the two equii join and natural join are equall.will both display even
output?Hello
It keeps you a little test and check yourself?See the link below.
http://psoug.org/reference/joins.html
Connected to: Oracle Database 11g Enterprise Edition Release 11.2.0.1.0 - Production With the Partitioning, OLAP, Data Mining and Real Application Testing options SQL> CREATE TABLE parents ( 2 person_id NUMBER(5), 3 adult_name VARCHAR2(20), 4 comments VARCHAR2(40)) 5 PCTFREE 0; Table created. SQL> SQL> CREATE TABLE children ( 2 parent_id NUMBER(5), 3 person_id NUMBER(5), 4 child_name VARCHAR2(20), 5 comments VARCHAR2(40)) 6 PCTFREE 0; Table created. SQL> SQL> INSERT INTO parents VALUES (1, 'Dan', 'So What'); 1 row created. SQL> INSERT INTO parents VALUES (2, 'Jack', 'Who Cares'); 1 row created. SQL> INSERT INTO children VALUES (1, 2, 'Anne', 'Who Cares'); 1 row created. SQL> INSERT INTO children VALUES (1, 1, 'Julia', 'Yeah Right'); 1 row created. SQL> INSERT INTO children VALUES (2, 1, 'Marcella', 'So What'); 1 row created. SQL> COMMIT; Commit complete. SQL> SQL> SELECT adult_name, child_name 2 FROM parents NATURAL JOIN children; ADULT_NAME CHILD_NAME -------------------- -------------------- Jack Anne Dan Marcella SQL> select adult_name,child_name from parents a, children b 2 where a.person_id=b.person_id; ADULT_NAME CHILD_NAME -------------------- -------------------- Jack Anne Dan Julia Dan Marcella SQL> ed Wrote file afiedt.buf 1 select adult_name,child_name from parents a, children b 2* where a.person_id=b.parent_id SQL> / ADULT_NAME CHILD_NAME -------------------- -------------------- Dan Anne Dan Julia Jack Marcella SQL>Kind regards
Avinash -
Hello
After reading a natural join on wikipedia , I was interested to use this syntax for expressing equi-joins between the tables that share some common column names. I have so these 3 tables:
-Courses (acronym, title, nbCredits)
-Registration (acronym, garnish, matricule, numSect, cumulative, noteFinale)
-Nobody (SIN, name, typPers, service number, pmatricule)
When I specify * in the select clause, I don't get that 3 lines, which is logical given the restriction I've specified in where clause. But what I don't understand is that if I replace the * with specific column names, then oracle returns a whole bunch of additional lines that should have been rejected in the where clause. It seems to me that Oracle should return only the next 3 rows. Does make sense to anyone? Here's the query I am running:
SELECT *.
OF course, NATURAL JOIN, NATURAL JOIN registration person
WHERE name = 'John '.
AND name = "Doe".
AND nbCredits > 2
AND noteFinale <>'F'
AND noteFinale IS NOT NULL;
-3 lines selected (right)
Individual SELECTION, the title
OF course, NATURAL JOIN, NATURAL JOIN registration person
WHERE name = 'John '.
AND name = "Doe".
AND nbCredits > 2
AND noteFinale <>'F'
AND noteFinale IS NOT NULL;
-80 (bad) selected lines
Ludovic
Published by: user3968717 on December 5, 2009 12:41
Published by: user3968717 on December 5, 2009 12:42I can't reproduce the problem - I get 1 row in both cases
SQL> ed Wrote file afiedt.buf 1 SELECT sigle, titre 2 FROM Cours NATURAL JOIN Inscription NATURAL JOIN Personne 3 WHERE prenom='Pierre' 4 AND nom='Daigneault' 5 AND nbCredits>2 6 AND noteFinale != 'F' 7* AND noteFinale IS NOT NULL SQL> / SIGLE ------------ TITRE -------------------------------------------------------------------------------- CY130 Clavardage et communications 1 row selected. SQL> ed Wrote file afiedt.buf 1 SELECT * 2 FROM Cours NATURAL JOIN Inscription NATURAL JOIN Personne 3 WHERE prenom='Pierre' 4 AND nom='Daigneault' 5 AND nbCredits>2 6 AND noteFinale != 'F' 7* AND noteFinale IS NOT NULL SQL> / MATRICULE SIGLE ---------- ------------ TITRE -------------------------------------------------------------------------------- NBCREDITS TRIM NUMSECT CUMULATIF NO NAS ---------- ---- ---------- ---------- -- ---------- NOM -------------------------------------------------------------------------------- PRENOM -------------------------------------------------------------------------------- T PMATRIC - ------- 1470609 CY130 ... MATRICULE SIGLE ---------- ------------ TITRE ------------------------------------------------------------------------------- NBCREDITS TRIM NUMSECT CUMULATIF NO NAS ---------- ---- ---------- ---------- -- ---------- NOM ------------------------------------------------------------------------------- PRENOM ------------------------------------------------------------------------------- T PMATRIC - ------- 1 row selected.I use 11.1.0.7, what version are you using? He had a number of bugs with joins SQL 99 when they were introduced. Maybe you're using an older version? Did you ask the most recent Group of patches for your version of Oracle?
I would caution, however, I would strongly, strongly suggest avoiding natural joins. In my mind, they are an abomination of SQL syntax. I love the syntax of SQL 99 and I'm perfectly happy with the syntax
x JOIN y USING (column_name )or
x JOIN y ON (x.column_name = y.column_name)But join based on all the identical column names causes a lot of problems in maintenance. With any other join syntax, if you si vous allez are going to add new columns to a table, the results of your SQL queries do not change. If someone decides that they must add columns to one of your tables in the future (CREATED_BY, LAST_MODIFIED_DATE, etc.), your queries suddenly start returning incorrect data. You waive any reasonable chance to make a reasonable assessment.
Justin
-
Hello!
I learn SQL and I have a lot of doubts but I think with this example I can generalize.
I have the tables:
Book {idbook (PK), namebook}
auhor {idauthor (PK), nameauthor, idcountry (FK)}
fatherhood {idauthor (FK), idbook (FK)} (both the same strain of PK)
country {idcountry (PK), namecountry}
I want the name of the books that are the authors of the Canada.
I assumed that a correct query would be:
SELECT namebook FROM book NATURAL JOIN (paternity NATURAL JOIN (author country NATURAL JOIN)) WHERE country.namecountry = 'Canada ';
The result I was waiting for was:
Book3
Book5
but this query returns me all the books that have relationships in the co-author (with the authors of all countries), 2 times! as:
Book2
book3
Book4
Book5
Book2
book3
Book4
Book5
the best I can do to get my correct result is:
SELECT namebook FROM book NATURAL JOIN (author, NATURAL JOIN) WHERE author.idcountry = 2;
But of course I can't use this one...
Can someone explain to me what is happening?
Thank you very much!
Published by: user12040235 on 10/15/2009 09:37
Published by: user12040235 on 10/15/2009 09:51Hello
Maybe it's a bug.
Get the correct results (2 ranks) to Oracle 10.1, but I get the same bad results you (12 rows) in Oracle 11.1.
In Oracle 11, I get results so say "SELECT *" instead of "SELECT book.namebook".
I also get the correct results, if I add the join columns in the SELECT clause. Adding a column to join no, for exampleSELECT BOOK.namebook, author.nameauthor FROM book ...Gets results.
In the interest of all those who want to try this:
DROP TABLE author; create table author AS SELECT 1 AS idauthor, 'Jose Luiz do Rego' AS nameauthor, 1 AS idcountry FROM dual UNION ALL SELECT 2, 'Barbara Bela', 2 FROM dual UNION ALL SELECT 3, 'Juan Domingues', 5 FROM dual UNION ALL SELECT 4, 'José Mauro de Vasconcelos', 1 FROM dual UNION ALL SELECT 5, 'Vader', 2 FROM dual UNION ALL SELECT 6, 'navathe', 4 FROM dual UNION ALL SELECT 7, 'Machado de Assis', 1 FROM dual ; drop table AUTHORSHIP; CREATE TABLE authorship AS SELECT 2 AS idauthor, 5 AS idbook FROM dual UNION ALL SELECT 1 AS idauthor, 1 AS idbook FROM dual UNION ALL SELECT 5 AS idauthor, 3 AS idbook FROM dual UNION ALL SELECT 6 AS idauthor, 2 AS idbook FROM dual UNION ALL SELECT 7 AS idauthor, 4 AS idbook FROM dual UNION ALL SELECT 7 AS idauthor, 6 AS idbook FROM dual; drop table book; CREATE TABLE book AS SELECT 1 AS idbook, 'book1' AS namebook FROM dual UNION ALL SELECT 2 AS idbook, 'book2' AS namebook FROM dual UNION ALL SELECT 3 AS idbook, 'book3' AS namebook FROM dual UNION ALL SELECT 4 AS idbook, 'book4' AS namebook FROM dual UNION ALL SELECT 5 AS idbook, 'book5' AS namebook FROM dual UNION ALL SELECT 6 AS idbook, 'book6' AS namebook FROM dual UNION ALL SELECT 7 AS idbook, 'book7' AS namebook FROM dual; DROP TABLE country; CREATE TABLE country AS SELECT 1 AS idcountry, 'Brazil' as namecountry FROM dual UNION ALL SELECT 2 AS idcountry, 'Canada' as namecountry FROM dual UNION ALL SELECT 3 AS idcountry, 'Chile' as namecountry FROM dual UNION ALL SELECT 4 AS idcountry, 'Venezuela' as namecountry FROM dual UNION ALL SELECT 5 AS idcountry, 'USA' as namecountry FROM dual UNION ALL SELECT 6 AS idcountry, 'Argentina' as namecountry FROM dual ; -
In the HR schema, to the natural join done on employees of three tables, different places, that no row is returned and Department based on the columns in the select list
For the query
Select country_id, location_id, department_id as well as wages of employees
departments of natural join
locations of natural join32 rows are to be extracted.
Time of the request
Select the salary, department_id from employee
departments of natural join
locations of natural join736 lines are retrieved.
Please explain how the name of column in select affect the natural join of the list
Hello
suchibm wrote:
Hi frank,.
Thanks for the advice, but it's accreditation. I use Oracle XE Databae.
But which version? If you are unsure, run
SELECT *.
SINCE the release of v$.
The important part will be the bunch of numbers towards the end of the 1st line.
This looks like a bug has been fixed in version 11.2.0.2.0 (or arose after that version).
A book of preparation of certification (or something like that), said you should get different results of these 2 queries?
If not, ignore the result 736 set line; 32 is the expected number of rows.
Just curious, when you get 736 lines, they are 23 copies of the same 32 ranks that you get in the query on the other? Use COUNT (*) in a GROUP BY query to find out.
It's a shame that certification documents include such useless things as NATURAL JOIN. While you are studying for certification, invest your time in something that you can use in real life. There must be other subjects that you don't fully understand and do not have bugs in your version.
-
Beginner help - faced with natural join as an example
Hello
My apologies for how basic this query probably will, but I am teaching from zero and do not like to move from topic to topic without a good understanding of why I get the results I get, rather than knowing 'it's how he did
I use the guide of the official examination for 1z0 - 051, and 320 page there is the following year:
+ "The table JOB_HISTORY sharing three columns with the same named with the EMPLOYEES table: EMPLOYEE_ID JOB_ID and department_id." You are required to describe the tables and fetch the EMPLOYEE_ID, JOB_ID, +.
+ Values for all rows retrieved using a natural join pure department_id, LAST_NAME, HIRE_DATE and end_date. Alias the table EMPLOYEES as EMP and the table JOB_HISTORY that JH and use dot notation where possible. » +
Their solution (which is about what I came to start with) is:
SELECT employee_id job_id, department_id, emp.last_name, emp.hire_date, JH.end_date
OF JH job_history
NATURAL JOIN employees emp;
This translates into the only employee "Taylor" being returned.
------------------------------
Is it possible someone could 'hold my hand' through this example please? I have no idea why only one result is returned, but in addition, I don't know what else I expected - clearly my understanding of the natural join clause is severely lacking. The book States that:
+ "The execution of this statement returns a single row with the same values EMPLOYEE_ID JOB_ID and department_id in the two tables." +
I guess I'm confused because I thought that the join clauses should deploy additional content to other tables to display, to not limit lines - which is what it sounds like they mean by "returns a single row with the same number and job_id department_id values in the two tables."
I am very confused!
Thanks in advance,
NickHi, Nick.
Nick woodward wrote:
HelloMy apologies for how basic this query probably will, but I am teaching from zero and do not like to move from topic to topic without a good understanding of why I get the results I get, rather than knowing 'it's how he did
Good thinking!
I use the guide of the official examination for 1z0 - 051, and 320 page there is the following year:
+ "The table JOB_HISTORY sharing three columns with the same named with the EMPLOYEES table: EMPLOYEE_ID JOB_ID and department_id." You are required to describe the tables and fetch the EMPLOYEE_ID, JOB_ID, +.
+ Values for all rows retrieved using a natural join pure department_id, LAST_NAME, HIRE_DATE and end_date. Alias the table EMPLOYEES as EMP and the table JOB_HISTORY that JH and use dot notation where possible. » +Their solution (which is about what I came to start with) is:
SELECT employee_id job_id, department_id, emp.last_name, emp.hire_date, JH.end_date
OF JH job_history
NATURAL JOIN employees emp;This translates into the only employee "Taylor" being returned.
------------------------------
Is it possible someone could 'hold my hand' through this example please? I have no idea why the only one result is returned,
This is because there is only one combination of rows from the tables emp and jh such as the join condition is TRUE. In a natural join, the join condition is not specified, so it would be better if you started looking at the equivalent natural join:
SELECT emp.employee_id , emp.job_id , emp.department_id , emp.last_name , emp.hire_date , jh.end_date FROM hr.job_history jh JOIN hr.employees emp ON emp.employee_id = jh.employee_id AND emp.job_id = jh.job_id AND emp.department_id = jh.department_id ;on the other hand I don't know what I was expecting - clearly my understanding of the natural join clause is severely lacking. The book States that:
+ "The execution of this statement returns a single row with the same values EMPLOYEE_ID JOB_ID and department_id in the two tables." +
I guess I'm confused because I thought that the join clauses should deploy additional content to other tables to display, do not limit the rows...
A join does both. Each row in the result set can have columns (or drift) the columns in the table and it can have any row of a table associated with the lines of all (or part) of the other table. There are 107 lines in the emp table and 10 rows in the table of jh. This means that could result in a join 2 tables can 0-107 * 10 = 1070 lines. If the join conditions are so strict that no combination of lines met, then the result set has 0 rows. However, if they are so loose that all combinations of lines 2 brands joins tables condition TRUE, then each of the 107 rows in the emp table will join each of the 10 lines in the table of jh.
Try to play with the join condition and see what happens. For example, we will comment on 2 of the 3 parts of the join condition, so that the only condition on the left is "emp.employee_id = jh.employee_id". Also, we'll include a few other columns in the table of jh and no poster not dates, just to make room for the new coplumns. So, if we execute this query:
SELECT emp.employee_id , emp.job_id , emp.department_id , emp.last_name -- , emp.hire_date -- Commented out to save space -- , jh.end_date -- Commented out to save space , jh.job_id -- For debugging , jh.department_id -- For debugging FROM hr.job_history jh JOIN hr.employees emp ON emp.employee_id = jh.employee_id -- AND emp.job_id = jh.job_id -- AND emp.department_id = jh.department_id ;then we get these results:
EMPLOYEE DEPARTMENT DEPARTMENT _ID JOB_ID _ID LAST_NAME JOB_ID _ID -------- ---------- ---------- ---------- ---------- ---------- 101 AD_VP 90 Kochhar AC_ACCOUNT 110 101 AD_VP 90 Kochhar AC_MGR 110 102 AD_VP 90 De Haan IT_PROG 60 114 PU_MAN 30 Raphaely ST_CLERK 50 122 ST_MAN 50 Kaufling ST_CLERK 50 176 SA_REP 80 Taylor SA_REP 80 176 SA_REP 80 Taylor SA_MAN 80 200 AD_ASST 10 Whalen AD_ASST 90 200 AD_ASST 10 Whalen AC_ACCOUNT 90 201 MK_MAN 20 Hartstein MK_REP 20 10 rows selected.Why do we get only 10 lines? There are other employee_ids, such as 100 and 206 in the emp table. Why do we see these employee_ids in the results? In addition, the emp.employee_id is unique; in other words, no 2 rows in the emp table have the same employe_id. How the result set can have two lines with the same employee_ids, such as 101 or 200?
Now try to do the more difficult condtion join, by a non-commenting on the terms that we have commented earlier:
SELECT emp.employee_id , emp.job_id , emp.department_id , emp.last_name -- , emp.hire_date -- , jh.end_date , jh.job_id , jh.department_id FROM hr.job_history jh JOIN hr.employees emp ON emp.employee_id = jh.employee_id AND emp.job_id = jh.job_id -- *** CHANGED *** -- AND emp.department_id = jh.department_id ;The results are now:
EMPLOYEE DEPARTMENT DEPARTMENT _ID JOB_ID _ID LAST_NAME JOB_ID _ID -------- ---------- ---------- ---------- ---------- ---------- 200 AD_ASST 10 Whalen AD_ASST 90 176 SA_REP 80 Taylor SA_REP 80He was 2 of 10 lines in the previous result set. What happened to the other 8 lines that have been in this result set? Fior example, why has this line:
EMPLOYEE DEPARTMENT DEPARTMENT _ID JOB_ID _ID LAST_NAME JOB_ID _ID -------- ---------- ---------- ---------- ---------- ---------- 176 SA_REP 80 Taylor SA_MAN 80in the result of 10 rows value (produced by the query with a join condition only 1), but not in the result of 2 ranks (produced by the query with the additional condition 'AND emp.job_id = jh.job_id')?
If you do not completely understand the natural joins, don't worry too much. In practice, nodody uses a natural join. However, natural joins are a type of INNER join and inner joins are extremely frequent. It is very important for you to understand how joins inside, such as the 2 I posted more top, work.
Published by: Frank Kulash, March 18, 2013 17:08
Results and applications added -
Select * from version of v$.
Oracle Database 11 g Enterprise Edition Release 11.2.0.1.0 - 64 bit Production
Why these constructions are analysed with success and full/cross/left are ignored?
Select * from
(select 1 of the double)
natural full join
(select 2 from two);
Select * from
(select 1 of the double)
natural left join
(select 2 from two);
Select * from
(select 1 of the double)
right to natural join
(select 2 from two);
Select * from
(select 1 of the double)
cross the natural join
(select 2 from two);
The same time if we try to alias tables it gets an error:
Select * from
(select double 1) t1
natural full join
(select double 2) t2;
ORA-00905: lack of keyword
00905 00000 - 'lack the key word'
* Cause:
* Action:
Error on line: column 5: 29
Published by: Slobodcicov on November 28, 2012 23:46Try this...
select * from (select 1 a from dual) t1 natural full join (select 2 a from dual) t2;A
--
2
1See you soon,.
Manik. -
Hi all I have a basic sql question
Watch below two querries
I think as operator will not compare with all the values, but not the outcome it will compare with all values... When comparing with the value null, the result is automatically null1. select 1 from dual where 1 in (select 1 from dual union all select null from dual) It gives output as 1 but below one 2. select 1 from dual where 1 not in (select 2 from dual union all select null from dual) It gives output as no data found
I'm wrong
Please help me on this
and why performance wise in operator is more better than no of?
Thanks to all in advanceThanks for posting your explain plan command
Execution Plan ---------------------------------------------------------- Plan hash value: 3249215828 ----------------------------------------------------------------------------- | Id | Operation | Name | Rows | Bytes | Cost (%CPU)| Time | ----------------------------------------------------------------------------- | 0 | SELECT STATEMENT | | 2 | 6 | 4 (0)| 00:00:01 | | 1 | NESTED LOOPS | | 2 | 6 | 4 (0)| 00:00:01 | | 2 | FAST DUAL | | 1 | | 2 (0)| 00:00:01 | | 3 | VIEW | VW_NSO_1 | 2 | 6 | 2 (0)| 00:00:01 | | 4 | SORT UNIQUE | | 2 | | 2 (0)| 00:00:01 | | 5 | UNION-ALL | | | | | | | 6 | FAST DUAL | | 1 | | 2 (0)| 00:00:01 | |* 7 | FILTER | | | | | | | 8 | FAST DUAL | | 1 | | 2 (0)| 00:00:01 | ----------------------------------------------------------------------------- Predicate Information (identified by operation id): --------------------------------------------------- 7 - filter(NULL IS NOT NULL) 02:12:54 SQL> select 1 from dual where 1 not in (select 2 from dual union all select null from dual); Elapsed: 00:00:00.01 Execution Plan ---------------------------------------------------------- Plan hash value: 3291682568 ----------------------------------------------------------------- | Id | Operation | Name | Rows | Cost (%CPU)| Time | ----------------------------------------------------------------- | 0 | SELECT STATEMENT | | 1 | 4 (0)| 00:00:01 | |* 1 | FILTER | | | | | | 2 | FAST DUAL | | 1 | 2 (0)| 00:00:01 | | 3 | UNION-ALL | | | | | |* 4 | FILTER | | | | | | 5 | FAST DUAL | | 1 | 2 (0)| 00:00:01 | | 6 | FAST DUAL | | 1 | 2 (0)| 00:00:01 | ----------------------------------------------------------------- -
A curious "anomaly" with the NATURAL JOIN compared to join HER for HELP
The following query returns I expect since the HR [106 lines] schema:
Select last_name, department_name of departments who JOIN employees USE (department_id);
However, when I do the full natural join with this, I get only 32 ranks.
Select last_name, department_name of the departments of NATURAL JOIN employees;
The complete NATURAL JOIN not use department_id join on? But if not what else would it use? I just stumbled on this and am especially curious that I do not use the syntax of natural join in production but prefer instead the JOIN now.
Thank you.The NATURAL keyword indicates that a natural join is underway. A natural join is based on all of the columns in the two tables with the same name. He selects the rows of two tables that have equal values in the corresponding columns.
-
Apex SQL question - weeks of the month
I have a sql question.
I want to create a dynamic list.
If the user select may2009, then I want a dynamic list of show
04/27/2009-05/03/2009
05/04/2009-05/10/2009
05/11/2009-05/17/2009
05/18/2009-05/24/2009
05/25/2009-05/31/2009
If the user selects Jun 2009, then the list will be
06/01/2009-06/07/2009
06/08/2009-06/14/2009
06/15/2009-06/21/2009
06/22/2009-06/28/2009
06/29/2009-07/05/2009
Thank you.Using this SQL statement, you can get this list:
SELECT w_start || ' - ' || w_end d, w_start r FROM (SELECT (week_start_list + (LEVEL - 1) * 7) + 1 w_start, week_start_list + (LEVEL) * 7 w_end FROM (SELECT TO_CHAR (TRUNC (TO_DATE (:my_date, 'dd.mm.yyyy'), 'mm' ), 'IW' ) week_begin, TRUNC (TO_DATE (:my_date, 'dd.mm.yyyy'), 'mm' ) m_begin, TO_CHAR (TRUNC (TO_DATE (:my_date, 'dd.mm.yyyy'), 'mm'), 'd' ) day_month_begin, TO_CHAR (TRUNC (ADD_MONTHS (TO_DATE (:my_date, 'dd.mm.yyyy' ), 1 ), 'mm' ), 'IW' ) week_end, TRUNC (ADD_MONTHS (TO_DATE (:my_date, 'dd.mm.yyyy'), 1 ), 'mm' ) m_end, TO_CHAR (TRUNC (ADD_MONTHS (TO_DATE (:my_date, 'dd.mm.yyyy' ), 1 ), 'mm' ), 'd' ) day_month_end, TRUNC (TO_DATE (:my_date, 'dd.mm.yyyy'), 'mm' ) - TO_NUMBER (TO_CHAR (TRUNC (TO_DATE (:my_date, 'dd.mm.yyyy' ), 'mm' ), 'd' ) ) week_start_list, TRUNC (ADD_MONTHS (TO_DATE (:my_date, 'dd.mm.yyyy'), 1 ), 'mm' ) + TO_NUMBER (TO_CHAR (TRUNC (ADD_MONTHS (TO_DATE (:my_date, 'dd.mm.yyyy' ), 1 ), 'mm' ), 'd' ) ) - 1 week_end_list FROM DUAL) CONNECT BY LEVEL <= (SELECT TO_NUMBER (TO_CHAR (TRUNC (ADD_MONTHS (TO_DATE (:my_date, 'dd.mm.yyyy' ), 1 ), 'mm' ), 'IW' ) ) - TO_NUMBER (TO_CHAR (TRUNC (TO_DATE (:my_date, 'dd.mm.yyyy' ), 'mm' ), 'IW' ) ) FROM DUAL))It is a question of SQL and has nothing to do with the Apex.
Denes Kubicek
-------------------------------------------------------------------
http://deneskubicek.blogspot.com/
http://www.Opal-consulting.de/training
http://Apex.Oracle.com/pls/OTN/f?p=31517:1
------------------------------------------------------------------- -
SQL * PLUS connects with SQL > question
People,
I'm confused on the command. / sqlplus with Oracle database.
According to my understanding,. / sqlplus connected with SQL > using option 3: sysdba, sysoper, sysasm.
I connect with SQL > to run the rel853.sql script to create a table PSOPRDEFN but a field "OPERPSWDSALT" did not appear in the table PSOPRDEFN although it is in the Create Table statement.
I connect with SQL * in the directory/home/user/OracleDB_Home/bin as below:
$ export SYSTEM_PASS = AccessId/mypass$ export ORACLE_HOME = / home/user/OracleDB_Home
$ export ORACLE_SID = HRCS90
$. / lsnrctl start LISTENER
$. / AccessId/mypass sqlplus as sysdba
SQL > startup
SQL > @/opt/PT8.53/scripts/rel853.sql
SQL > select OPERPSWDSALT in the AccessId.PSOPRDEFN;
It returns: "OPERPSWDSALT": invalid identifier.
I checked the table PSOPRDEFN that the OPERPSWDSALT field did not appear in the table PSOPRDEFN.I tried the sysoper option as below:
$. / sqlplus AccessId/mypass as sysoper
SQL > startup
SQL > @/opt/PT8.53/scripts/rel853.sql
SQL > select OPERPSWDSALT in the AccessId.PSOPRDEFN;
It returns: "OPERPSWDSALT": invalid identifier.
The sysoper error is the same thing with sysdba error.
I tried the sysasm option as below:
$. / sqlplus AccessId/mypass as sysasm
He returned: connection refused.
Someone told me this connection as SYS causes this error. If not a sysdba, sysoper, or sysasm, use the command. / sqlplus AccessId/mypass cannot connect with SQL >.
My question is:
First of all, why is what OPERPSWDSALT did not appear in the PSOPRDEFN table while it is in the CREATE TABLE statement?
Seocond, if not use SYS which is one of the 3 options, how to run the command. / sqlplus AccessId/mypass to connect with SQL > so that CREATE TABLE PSOPRDEFN correctly?Thank you.
user8860348 wrote:
People,
Hello. Thanks much for the reply. I just do the commands below:
$ export SYSTEM_PASS = AccessId/mypass
$ export ORACLE_HOME = / home/user/OracleDB_Home
$ export ORACLE_SID = HRCS90
$. / lsnrctl start LISTENER
$. / AccessId/mypass sqlplus as sysdba
SQL > show user;
Its release: the USER is "SYS".
SQL > connect AccessId/mypass
Its output:
Error: ORA - 01034:ORACLE not available
ORA-27101: shared memory realm does not exist
64 - Linux_x86 error: no such file or directory.
As we see above, unable to connect to the Oracle database AccessId.
My question is:
What to do on AccessId, so that it can connect to the Oracle database?
Thank you.
so much for the use that you refuse to actually use COPY it PASTE & so that we can see the whole session.
In the past, you did
> SQL > startup
Maybe the database is out of order & must be started.
do exactly as below (line by line)
ID
sqlplus
/ as sysdba
startup
connect AccessId/mypass
COPY the results from above then PASTE all back here
-
SQL question: horizontally to group by
Hello
I have a problem to find a way to display the result of a join table. Can someone please help?
SQL > select * from tab1;
GNAME SNO
---------- ----------
ABC 5
DEF 2
ABC 3
SQL > select * from tab2.
GNAME SNO
---------- ----------
ABC 10
XYZ 5
DEF 15
My output using a UNION ALL operation current:
SQL > select gname, sno1 count (*) Group of tab1 by gname
2 Union all the
3 select gname, sno2 count (*) of the tab2 group by gname.
SNO1 GNAME
---------- ----------
ABC 2
DEF 1
ABC 1
DEF 1
XYZ 1
The output I expect:
GNAME SNO1 SNO2
---------- ---------- ----------
1 2 ABC
1 1 DEF
0 1 XYZ
Thanks in advance.Way the easiest would be to encapsulate a slightly modified version of your query in an external as query:
select gname, sum(sno1) sno1, sum(sno2) sno2 from (select gname, count(*) sno1 0 sno2 from tab1 group by gname union all select gname, 0 sno1, count(*) sno2 from tab2 group by gname);John
-
Hi expert,
When I ran after SQL, error message reads "table or view does not exist" pointing to the table "dba_tab_cols" and "dba_all_tables". There is no other question for this statement, because if I changed table "user_tab_cols" and "user_all_tables", it works well.
declare
v_old_table DBA_tab_columns.table_name%type;
v_where Varchar2 (4000);
Boolean v_first_col: = true;
type rc is ref cursor;
c rc;
v_rowid varchar2 (20);
Val varchar2 (50): = "Test note";
Start
for r in)
Select
t.*
Of
dba_tab_cols t, dba_all_tables a
where t.table_name = a.table_name
and t.data_type like '% CHAR % '.
and a.owner = 'QA'
order by t.table_name loop)
If v_old_table is null then
v_old_table: = r.table_name;
end if;
If v_old_table <>r.table_name then
v_first_col: = true;
-dbms_output.put_line ('search' | v_old_table);
Open c for ' select rowid from ' ' |. ' v_old_table | '" ' || v_where;
extract the c in v_rowid;
loop
When the output c % notfound;
dbms_output.put_line (' rowid: ' | v_rowid |) "in" | v_old_table);
extract the c in v_rowid;
end loop;
v_old_table: = r.table_name;
end if;
If v_first_col then
v_where: = 'where ' | r.column_name | "as" %' | Val | '%''';
v_first_col: = false;
on the other
v_where: = v_where | "or" | r.column_name | "as" %' | Val | '%''';
end if;
end loop;
end;
But if I choose these DBA tables in the toad sql Editor, it works well, I am currently using my own credentials, not the administrator credentials. why he get different effects running in these two ways?
Thank you very much>
Hello
When I ran after SQL, error message reads "table or view does not exist" pointing to
Table 'dba_tab_cols' and 'dba_all_tables '. There is no other question for this statement, becauseIf I changed table "user_tab_cols" and "user_all_tables", it works well.
But if I choose these DBA tables in the toad sql Editor, it works well, I am currently using my own
credentials, not the administrator credentials. why he get different effects running in these two ways?You answered your own question - you know that you have the administrator privileges when you
Open a session under your own credentials - but your id user obviously does TOAD.You use SQL * for the query that fails?
BTW, you do not give us your version of Oracle - you must always tell us what it is
Please read the forum FAQ and also the thread "sticky" by BluShadow at the top of the list of positions
on the forum homepage. These forums are an excellent resource - you will get the best out of them if
you follow the instructions.HTH,
Paul...
Published by: Paulie July 24, 2012 16:40
Maybe you are looking for
-
When I call my wife on facetime (on his Apple ID, iPhone 5 iOS 7. XX), there is another person who answered my call, that I don't even know. I called 3 times my wife between yesterday and today, and the same guy responded twice. It's really strange..
-
Video capture without tuner TV Qosmio F10
Hello I have a Qosmio F10-101. It does not include a TV tuner, but I have composite and s video in. How can I capture any film?Please I really need help!
-
I want to edit/change my username to my new computer. How can I do?
I'm in the middle of my new Omni 120 PC computer programming (APU AMD-E-450 with graphics HD Radeon (TM)), but I want to change the username (you mind, not add). How can I do? I still have to install the anti-virus protection on it.
-
My win xp sp3 displays this error 0X000000F4 (0 x 00000003, 0X80ED4020, 0X80ED4194, 0X805D2970) in the start time menu some time pc works without error after 2 or 3 p.m. shows this message what solution help me please
-
Stupid question about the signing of applications made in Cascade
Hello, I am sure that it is on the site somewhere and I'll be stupid. But how sign and get the files and is not so you can put a cascade of the project on the app world? Also its for blackberry 10 Thank you