How to get the parent given a child in SQL SERVER 2005 - sql-server-2005

I have a table like this
childid parentid
------------------------
1 0
2 1
3 2
4 2
5 3
6 4
7 0
8 7
9 8
10 1
If I give a childid as 5, the parentid will be 1(output)
If I give a childid as 9, the parentid will be 7.(output)
i.e. the root parentid is 0 and the query should stop there.
How to solve such a query?
Please help.

I think you should rename your child_id to node, your parent_id to child_of. Your column naming is a bit confusing
create table stack_overflow
(
node int, child_of int
);
insert into stack_overflow(node, child_of) values
(1,0),
(2,1),
(3,2),
(4,2),
(5,3),
(6,4),
(7,0),
(8,7),
(9,8),
(10,1);
This works on any CTE-capable RDBMS:
with find_parent(parent, child_of, recentness) as
(
select node, child_of, 0
from stack_overflow
where node = 9
union all
select i.node, i.child_of, fp.recentness + 1
from stack_overflow i
join find_parent fp on i.node = fp.child_of
)
select top 1 parent from find_parent
order by recentness desc
Output:
parent
7
[EDIT: more flexible and future-proof]:
with find_parent(node_group, parent, child_of, recentness) as
(
select node, node, child_of, 0
from stack_overflow
where node in (5,9)
union all
select fp.node_group, i.node, i.child_of, fp.recentness + 1
from stack_overflow i
join find_parent fp on i.node = fp.child_of
)
select q.node_group as to_find, parent as found
from find_parent q
join
(
select node_group, max(recentness) as answer
from find_parent
group by node_group
) as ans on q.node_group = ans.node_group and q.recentness = ans.answer
order by to_find
Output:
to_find found
5 1
9 7
If you're using Postgres, the above code could be shortened to:
with recursive find_parent(node_group, parent, child_of, recentness) as
(
select node, node, child_of, 0
from stack_overflow
where node in (5,9)
union all
select fp.node_group, i.node, i.child_of, fp.recentness + 1
from stack_overflow i
join find_parent fp on i.node = fp.child_of
)
select distinct on (node_group) node_group as to_find, parent as found
from find_parent
order by to_find, recentness desc
DISTINCT ON rocks! :-)

If ALL you want is the root ParentID, you can use this recursive function:
CREATE FUNCTION test_func
(
#ParentID int
)
RETURNS int
AS
BEGIN
DECLARE #result int;
DECLARE #childID int;
SET #childID = (SELECT ChildID FROM YourTable WHERE ParentID = #ParentID)
IF (#childID = 0)
SET #result = #ParentID
ELSE
SET #result = dbo.test_func(#childID)
RETURN #result
END
GO
then in your main query:
SELECT dbo.test_func(5)
Passing in 5 returns 1, 9 returns 7 based on your provided data. If you need every ParentID that is up that chain, you should probably use a CTE.

I think you want a recursive query, you should use Common Table Expressions. I will give you a link with an example very similar that the one you're using.
I think here is the solution. It helped me some months ago.

A simple of example of getting the parent ID matching a given child ID is:
select parentid
from MyTable
where childid = 5
However, for the data above, this will return no records.

Related

Getting different Hierarchy levels and Path in Snowflake SQL

I am trying to replicate the result of a qlik function called Hierarchy. It creates the Hierarchy with different Levels and also gives the Hierarchy Path. The Code that i am using so far is giving me the Levels but with an error that Level2 values are also coming in Level1.
CREATE OR REPLACE TRANSIENT TABLE "Hierarchy" ( "NodeID" VARCHAR, "ParentID" VARCHAR, "NodeName" String)
Insert into "Hierarchy"("NodeID", "ParentID","NodeName")
VALUES
('1','4','London'),
('2','3','Munich'),
('3','5','Germany'),
('4','5','UK'),
('5','', 'Europe');
with recursive CteTree as
(Select "NodeID","ParentID","NodeName" as "NodeName1",
CAST (NULL AS varchar(255)) as "NodeName2",
CAST (NULL AS varchar(255)) as "NodeName3",
0 as NodeName
from "Hierarchy"
Where "ParentID" is not null
UNION ALL
Select child."NodeID", child."ParentID", "NodeName1",
Case When NodeName+1 = 1 then "NodeName" else "NodeName2" end,
Case When NodeName+1 = 2 then "NodeName" else "NodeName3" end,
NodeName+1
from CteTree
Join "Hierarchy" child
ON child."ParentID" = CteTree."NodeID"
)
select distinct * from CteTree order by "NodeName1"
The Output that it is producing:
Desired OutPut:
How can it be achieved?
with Hierarchy (NodeID, ParentID,NodeName) as (select * from VALUES
('1','4','London'),
('2','3','Munich'),
('3','5','Germany'),
('4','5','UK'),
('5','', 'Europe'))
select
sys_connect_by_path(NodeName, ' -> ') path
, NodeID
, ParentID
, NodeName
from
Hierarchy
START WITH ParentID =''
CONNECT BY PRIOR NodeID = ParentID ;
CREATE TABLE HIERARCHYWITHLEVEL(NODEID, PARENTID, NODENAME, LEVEL)
AS
WITH TREE AS
(SELECT NODEID, PARENTID, NODENAME, 1 AS LEVEL
FROM HIERARCHY
WHERE PARENTID = ''
UNION ALL
SELECT HIERARCHY.NODEID, HIERARCHY.PARENTID, HIERARCHY.NODENAME, LEVEL + 1
FROM HIERARCHY
JOIN TREE
ON HIERARCHY.PARENTID = TREE.NODEID
)
SELECT NODEID, PARENTID, NODENAME, LEVEL
FROM TREE;
NODEID
PARENTID
NODENAME
LEVEL
5
Europe
1
3
5
Germany
2
4
5
UK
2
2
3
Munich
3
1
4
London
3

Finding missing values from a sequence in a SELECT statement [duplicate]

I have table with a unique auto-incremental primary key. Over time, entries may be deleted from the table, so there are "holes" in this field's values. For example, table data may be as follows:
ID | Value | More fields...
---------------------------------
2 | Cat | ...
3 | Fish | ...
6 | Dog | ...
7 | Aardvark | ...
9 | Owl | ...
10 | Pig | ...
11 | Badger | ...
15 | Mongoose | ...
19 | Ferret | ...
I'm interested in a query that will return the list of missing IDs in the table. For the data above, the expected results are:
ID
----
1
4
5
8
12
13
14
16
17
18
Notes:
It is assumed that the initial first ID was 1
The maximum ID that should be examined is the final one, i.e. it's okay to assume that there were no additional entries after the current last one (see additional data on this point below)
A drawback of the above requirements is that the list will not return IDs that were created after ID 19 and that were deleted. I'm currently solving this case in code, because I hold the max ID created. However, if the query can take as a parameter MaxID, and also return those IDs between the current max and MaxID, that would be a nice "bonus" (but certainly not a must).
I'm currently working with MySQL, but consider moving to SQL Server, so I would like the query to fit both. Also, if you are using anything that can't run on SQLite, please mention it, thanks.
I landed on this page hoping to find a solution for SQLITE as this was the only answer I found when searching for this same question for SQLITE.
The final solution I found was from this article here
Float Middle Blog - SQLITE answer
Hope it helps someone else out :-)
the simple solution being:
SELECT DISTINCT id +1
FROM mytable
WHERE id + 1 NOT IN (SELECT DISTINCT id FROM mytable);
genius.
This question often comes up, and sadly, the most common (and most portable) answer is to create a temporary table to hold the IDs that should be there, and do a left join. The syntax is pretty similar between MySQL and SQL Server. The only real difference is the temporary tables syntax.
In MySQL:
declare #id int
declare #maxid int
set #id = 1
select #maxid = max(id) from tbl
create temporary table IDSeq
(
id int
)
while #id < #maxid
begin
insert into IDSeq values(#id)
set #id = #id + 1
end
select
s.id
from
idseq s
left join tbl t on
s.id = t.id
where t.id is null
drop table IDSeq
In SQL Server:
declare #id int
declare #maxid int
set #id = 1
select #maxid = max(id) from tbl
create table #IDSeq
(
id int
)
while #id < #maxid --whatever you max is
begin
insert into #IDSeq values(#id)
set #id = #id + 1
end
select
s.id
from
#idseq s
left join tbl t on
s.id = t.id
where t.id is null
drop table #IDSeq
Here's the query for SQL Server:
;WITH Missing (missnum, maxid)
AS
(
SELECT 1 AS missnum, (select max(id) from #TT)
UNION ALL
SELECT missnum + 1, maxid FROM Missing
WHERE missnum < maxid
)
SELECT missnum
FROM Missing
LEFT OUTER JOIN #TT tt on tt.id = Missing.missnum
WHERE tt.id is NULL
OPTION (MAXRECURSION 0);
Hope this is helpful.
PostgreSQL-only, inspired by other answers here.
SELECT all_ids AS missing_ids
FROM generate_series((SELECT MIN(id) FROM your_table), (SELECT MAX(id) FROM your_table)) all_ids
EXCEPT
SELECT id FROM your_table
I know it's an old question and already has an accepted answer,
but using a temp table isn't really necessary. Fixed formatting (sorry for double post).
DECLARE #TEST_ID integer, #LAST_ID integer, #ID integer
SET #TEST_ID = 1 -- start compare with this ID
SET #LAST_ID = 100 -- end compare with this ID
WHILE #TEST_ID <= #LAST_ID
BEGIN
SELECT #ID = (SELECT <column> FROM <table> WHERE <column> = #TEST_ID)
IF #ID IS NULL
BEGIN
PRINT 'Missing ID: ' + CAST(#TEST_ID AS VARCHAR(10))
END
SET #TEST_ID = #TEST_ID + 1
END
This is an Oracle only solution. It doesn't address the full question, but is left here for others that may be using Oracle.
select level id -- generate 1 .. 19
from dual
connect by level <= 19
minus -- remove from that set
select id -- everything that is currently in the
from table -- actual table
to get the missing rows from table
DECLARE #MaxID INT = (SELECT MAX(ID) FROM TABLE1)
SELECT SeqID AS MissingSeqID
FROM (SELECT ROW_NUMBER() OVER (ORDER BY column_id) SeqID from sys.columns) LkUp
LEFT JOIN dbo.TABLE1 t ON t.ID = LkUp.SeqID
WHERE t.ID is null and SeqID < #MaxID
I just have found the solution for Postgres:
select min(gs)
from generate_series(1, 1999) as gs
where gs not in (select id from mytable)
The single query can find the missing IDs..
SELECT distinct number
FROM master..spt_values
WHERE number BETWEEN 1 and (SELECT max(id) FROM MyTable)
AND number NOT IN (SELECT id FROM MyTable)
Update: This method took way too long so I wrote a linux command to find gaps in a text file. It does so in reverse order so first dump all id's to a text file like so;
nohup mysql --password=xx -e 'select id from tablename order by id desc' databasename > /home/ids.txt &
The first and last two lines are just to keep track of how long it took. 1.5million IDs(ish) took me 57sec & that's on a slow server. Set the max id in i and have at it.
T="$(date +%s)"; \
i=1574115; \
while read line; do \
if [[ "$line" != "$i" ]] ; then \
if [[ $i -lt 1 ]] ; then break; fi; \
if [[ $line -gt 1 ]] ; then \
missingsequenceend=$(( $line + 1 )); \
minusstr="-"; \
missingsequence="$missingsequenceend$minusstr$i"; \
expectnext=$(( $line - 1 )); \
i=$expectnext; \
echo -e "$missingsequence"; \
fi; \
else \
i=$(( $i - 1 )); \
fi; \
done \
< /home/ids.txt; \
T="$(($(date +%s)-T))"; \
echo "Time in seconds: ${T}"
Example output:
1494505-1494507
47566-47572
Time in seconds: 57
Also, I got syntax errors with the code from Eric's answer, but after changing the delimiter, using semicolons in the proper places and storing it in a procedure, it works.
Make sure you set the proper max ID, database name and table name (it's in the select query). And if you want to change the procedure name, change it in all 3 places.
use dbname;
drop procedure if exists dorepeat;
delimiter #
CREATE PROCEDURE dorepeat()
BEGIN
set #id = 1;
set #maxid = 1573736;
drop table if exists IDSeq;
create temporary table IDSeq
(
id int
);
WHILE #id < #maxid DO
insert into IDSeq values(#id);
set #id = #id + 1;
END WHILE;
select
s.id
from
IDSeq s
left join tablename t on
s.id = t.id
where t.id is null;
drop table if exists IDSeq;
END#
delimiter ;
CALL dorepeat;
I also found this query elwhere, but I haven't tested it.
SELECT a.id+1 AS start, MIN(b.id) - 1 AS end
FROM tablename AS a, tablename AS b
WHERE a.id < b.id
GROUP BY a.id
HAVING start < MIN(b.id)
TRY in MySQL
DELIMITER ||
DROP PROCEDURE IF EXISTS proc_missing ||
CREATE PROCEDURE proc_missing()
BEGIN
SET #minID = (SELECT MIN(`id`) FROM `tbl_name` WHERE `user_id`=13);
SET #maxID = (SELECT MAX(`id`) FROM `tbl_name` WHERE `user_id`=13);
REPEAT
SET #tableID = (SELECT `id` FROM `tbl_name` WHERE `id` = #minID);
IF (#tableID IS NULL) THEN
INSERT INTO temp_missing SET `missing_id` = #tableID;
END IF;
SET #minID = #minID + 1;
UNTIL(#minID <= #maxID)
END REPEAT;
END ||
DELIMITER ;
A few days ago, I was working on a production report and found some numbers missing. The missing numbers are very important, so I was asked to find a list of all missing numbers for investigation purposes. I posted a blog entry here, with a full demo, including a script to find missing numbers/IDs in a sample table.
The script suggested is quite long, so I won't include it here. Here are the basic steps used:
Create one temp table and store all distinct Numbers.
Find NextID which has something missing before it. Store into one TempTable.
Create one temp table to store missing number details.
Start to find the missing id using WHILE Loop.
Select missing data from #MissingID temp table.
Converting the SQL CTE (from Paul Svirin) to the Oracle version it looks like this (replace :YOURTABLE with the name of your table):
WITH Missing (missnum,maxid) as (
SELECT 1 missnum, (select max(id) from :YOURTABLE) maxid from dual
UNION ALL
SELECT m.missnum + 1,m.maxid
FROM Missing m
WHERE m.missnum < m.maxid
)
SELECT missnum
FROM Missing
LEFT OUTER JOIN :YOURTABLE tt on tt.id = Missing.missnum
WHERE tt.id is NULL
Using #PaulSvirin's answer, I've expanded it with a UNION to show ALL the data in my table, including the missing records with NULLs.
WITH Missing(missnum, maxid) AS
(SELECT (SELECT MIN(tmMIN.TETmeetingID)
FROM tblTETMeeting AS tmMIN)
AS missnum,
(SELECT MAX(tmMAX.TETmeetingID)
FROM tblTETMeeting AS tmMAX)
AS maxid
UNION ALL
SELECT missnum + 1, maxid
FROM Missing
WHERE missnum < maxid)
SELECT missnum AS TETmeetingID,
tt.DateID,
tt.WeekNo,
tt.TETID
FROM Missing LEFT JOIN tblTETMeeting tt ON tt.TETmeetingID = Missing.missnum
WHERE tt.TETmeetingID IS NULL
UNION
SELECT tt.TETmeetingID,
tt.DateID,
tt.WeekNo,
tt.TETID
FROM tblTETMeeting AS tt
OPTION ( MAXRECURSION 0 )
Work's great!
TETmeetingID DateID WeekNo TETID
29 3063 21 1
30 null null null
31 null null null
32 null null null
33 null null null
34 3070 22 1
35 3073 23 1
Easiest solution for me: Create a select that gives all ids up to max sequence value (ex:1000000), and filter:
with listids as (
Select Rownum idnumber From dual Connect By Rownum <= 1000000)
select * from listids
where idnumber not in (select id from table where id <=1000000)
A modified version borrowing #Eric proposal. This is for SQL Server and holds in a temp table the start and end value for missing ranges. If the gap is just one value it puts NULL as end value for easier visualization.
It will produce an output like this
|StartId| EndId |
|-------|-------|
| 1 | 10182 |
| 10189 | NULL |
| 10246 | 15000 |
And this is the script where myTable and id needs to be replaced by your table and identity column.
declare #id bigint
declare #endId bigint
declare #maxid bigint
declare #previousid bigint=0
set #id = 1
select #maxid = max(id) from myTable
create table #IDGaps
(
startId bigint,
endId bigint
)
while #id < #maxid
begin
if NOT EXISTS(select id from myTable where id=#id)
BEGIN
SET #previousid=#id
select top 1 #endId=id from myTable where id>#id
IF #id=#endId-1
insert into #IDGaps values(#id,null)
ELSE
insert into #IDGaps values(#id,#endId-1)
SET #id=#endId
END
ELSE
set #id = #id + 1
end
select * from #IDGaps
drop table #IDGaps
SOLUTION FOR SQLITE
if your table id only support positive values you can use this
SELECT DISTINCT table_id - 1 AS next_id
FROM table
WHERE next_id NOT IN (SELECT DISTINCT table_id FROM table)
AND next_id > 0
otherwise you should remove ids greater than the biggest id with
SELECT DISTINCT table_id + 1 AS next_id
FROM table
WHERE next_id NOT IN (SELECT DISTINCT table_id FROM table)
AND id < (SELECT MAX(id) FROM table)
This problem can be solved with only one query
select lft.id + 1 as missing_ids
from tbl as lft left outer join tbl as rght on lft.id + 1 = rght.id
where rght.id is null and lft.id between 1 and (Select max(id)-1 from tbl)
Tested on Mysql
Try This Query. This single query is enough to get missing numbers:(Please replace TABLE_NAME to which table name you are using)
select sno as missing from(SELECT #row := #row + 1 as sno FROM
(select 0 union all select 1 union all select 3 union all select 4 union all
select 5 union all select 6 union all select 6 union all select 7 union all
select 8 union all select 9) t,(select 0 union all select 1 union all select 3
union all select 4 union all select 5 union all select 6 union all select 6
union all select 7 union all select 8 union all select 9) t2,(select 0
union all select 1 union all select 3 union all select 4 union all select 5
union all select 6 union all select 6 union all select 7 union all select 8
union all select 9) t3, (select 0 union all select 1 union all select 3 union
all select 4 union all select 5 union all select 6 union all select 6 union all
select 7 union all select 8 union all select 9) t4,
(SELECT #row:=0) as b where #row<1000) as a where a.sno not in
(select distinct b.no from
(select b.*,if(#mn=0,#mn:=b.no,#mn) as min,(#mx:=b.no) as max from
(select ID as no from TABLE_NAME as a) as b,
(select #mn:=0,#mx:=0) as x order by no) as b) and
a.sno between #mn and #mx;
SELECT DISTINCT id -1
FROM users
WHERE id != 1 AND id - 1 NOT IN (SELECT DISTINCT id FROM users)
Explanation: ( id - 1 )..... checking for any previous id present in table
( id != 1 ).....neglecting when current id is 1 as its previous id will be 0 zero.
I have a large audit table and needed something that ran quickly - this worked well for me. It merges the top and bottom IDs for the missing ranges
select minQ.num,minId,maxId from
(SELECT DISTINCT id +1 as minId, Row_Number() Over ( Order By id ) As Num
FROM tblAuditLoghistory
WHERE id + 1 NOT IN (SELECT DISTINCT id FROM tblAuditLogHistory)
AND id < (SELECT max(id) FROM tblAuditLoghistory)) Minq
join
(SELECT DISTINCT id - 1 as maxId, Row_Number() Over ( Order By id ) As Num
FROM tblAuditLoghistory
WHERE id - 1 NOT IN (SELECT DISTINCT id FROM tblAuditLogHistory)
AND id > (SELECT min(id) FROM tblAuditLoghistory)) maxQ on minQ.num=maxQ.num
This what i used to find the missing id of one table named as tablename
select a.id+1 missing_ID from tablename a
where a.id+1 not in (select id from tablename b where b.id=a.id+1)
and a.id!=(select id from tablename c order by id desc limit 1)
It will return the missing ids.
If there are two(2) or more continuous missing ids, it will return only the first.

SQL Recursion: Get the main state off of a sub state

I found this nice example of SQL recursion with a CTE, but fail to apply it to my table:
http://walkingoncoals.blogspot.de/2009/12/fun-with-recursive-sql-part-1.html
I have the following table (ObjectStates):
ID Title ParentID
1 Draft null
2 Green null
3 Red null
4 Foo 1
5 Bar 4
I am trying to create a function which returns the "main" state when queried. Example:
GetMainState(5)
-- Shall return 1
GetMainState(4)
-- Shall return 1
GetMainState(2)
-- Shall return 2
I have so far:
CREATE FUNCTION [dbo].[GetMainObjectState] (#ObjectStateID INT)
RETURNS TABLE
AS
RETURN
(
WITH StateRecurcsion(ID, ParentID, Level) AS
(
SELECT ID, ParentID, 0
FROM ObjectStates
WHERE ID = #ObjectStateID
UNION ALL
SELECT uOS.ID, uOS.ParentID, sOS.Level+1
FROM ObjectStates uOS, StateRecurcsion sOS
WHERE uOS.ParentID= sOS.ID
)
SELECT os.ID, os.Title, sos.Level
FROM ObjectStates os, StateRecurcsion sos
WHERE os.ID = sos.ID
)
GO
I tried to create the function just as in the tutorial shown above, but somehow I'm not getting the correct results.
You could create a CTE containing a "root" value and then query it within your function e.g.:
;WITH CTEHierarchy
AS (
SELECT
ID
,0 AS LEVEL
,ID AS root
FROM ObjectStates
WHERE ParentID IS NULL
UNION ALL
SELECT
ObjectStates.ID
,LEVEL + 1 AS LEVEL
,[root]
FROM ObjectStates
INNER JOIN CTEHierarchy uh ON uh.id = ObjectStates.ParentID
)
SELECT [root]
FROM CTEHierarchy
WHERE ID = #ObjectStateID

sql select puzzle: remove children when parent is filtered out

I have a table essentially:
name has_children parent_id row_id values0.....valuesn
parent 1 1 1
children 0 1 2
children 0 1 3
parent 0 4 4
parent 1 5 5
children 0 5 6
children 0 5 7
the values for the children can be different than the values for the parent.
i want some selects/joins that will filter the table on a value column (i.e. >10) and will return the parent (even if false for the filter) if one of it's children is true for the filter.
acceptable return:
parent=true all children=false, return just parent
parent=false >=1 children=true, return parent and all non-filtered child
i'm sure this has been thought about before but i don't have the faintest idea how to phrase the question to find a solution.
ANSI compliant. Each specific DBMS may have a faster implementation
select *
from tbl
where id in-- PARENTS of CHILDREN that match
( select parent_id from tbl
where values0 > 10 and has_children = 0)
or id in -- ONE CHILD ONLY
( select MIN(id) from tbl
where values0 > 10 and has_children = 0
group by parent_id)
or id in -- PARENTS
( select id from tbl
where values0 > 10 and has_children = 1)
Better written as a JOIN
select t.*
from
( select parent_id as ID from tbl
where values0 > 10 and has_children = 0
UNION
select MIN(id) from tbl
where values0 > 10 and has_children = 0
group by parent_id
UNION
select id from tbl
where values0 > 10 and has_children = 1) X
join tbl t on X.ID = t.ID
It is probably easiest to deal with this as two separate queries with a UNION between them.
Select the parent where all children are false under condition.
Select the parent and some child (MAX or MIN is probably easiest).

SQL: find missing IDs in a table

I have table with a unique auto-incremental primary key. Over time, entries may be deleted from the table, so there are "holes" in this field's values. For example, table data may be as follows:
ID | Value | More fields...
---------------------------------
2 | Cat | ...
3 | Fish | ...
6 | Dog | ...
7 | Aardvark | ...
9 | Owl | ...
10 | Pig | ...
11 | Badger | ...
15 | Mongoose | ...
19 | Ferret | ...
I'm interested in a query that will return the list of missing IDs in the table. For the data above, the expected results are:
ID
----
1
4
5
8
12
13
14
16
17
18
Notes:
It is assumed that the initial first ID was 1
The maximum ID that should be examined is the final one, i.e. it's okay to assume that there were no additional entries after the current last one (see additional data on this point below)
A drawback of the above requirements is that the list will not return IDs that were created after ID 19 and that were deleted. I'm currently solving this case in code, because I hold the max ID created. However, if the query can take as a parameter MaxID, and also return those IDs between the current max and MaxID, that would be a nice "bonus" (but certainly not a must).
I'm currently working with MySQL, but consider moving to SQL Server, so I would like the query to fit both. Also, if you are using anything that can't run on SQLite, please mention it, thanks.
I landed on this page hoping to find a solution for SQLITE as this was the only answer I found when searching for this same question for SQLITE.
The final solution I found was from this article here
Float Middle Blog - SQLITE answer
Hope it helps someone else out :-)
the simple solution being:
SELECT DISTINCT id +1
FROM mytable
WHERE id + 1 NOT IN (SELECT DISTINCT id FROM mytable);
genius.
This question often comes up, and sadly, the most common (and most portable) answer is to create a temporary table to hold the IDs that should be there, and do a left join. The syntax is pretty similar between MySQL and SQL Server. The only real difference is the temporary tables syntax.
In MySQL:
declare #id int
declare #maxid int
set #id = 1
select #maxid = max(id) from tbl
create temporary table IDSeq
(
id int
)
while #id < #maxid
begin
insert into IDSeq values(#id)
set #id = #id + 1
end
select
s.id
from
idseq s
left join tbl t on
s.id = t.id
where t.id is null
drop table IDSeq
In SQL Server:
declare #id int
declare #maxid int
set #id = 1
select #maxid = max(id) from tbl
create table #IDSeq
(
id int
)
while #id < #maxid --whatever you max is
begin
insert into #IDSeq values(#id)
set #id = #id + 1
end
select
s.id
from
#idseq s
left join tbl t on
s.id = t.id
where t.id is null
drop table #IDSeq
Here's the query for SQL Server:
;WITH Missing (missnum, maxid)
AS
(
SELECT 1 AS missnum, (select max(id) from #TT)
UNION ALL
SELECT missnum + 1, maxid FROM Missing
WHERE missnum < maxid
)
SELECT missnum
FROM Missing
LEFT OUTER JOIN #TT tt on tt.id = Missing.missnum
WHERE tt.id is NULL
OPTION (MAXRECURSION 0);
Hope this is helpful.
PostgreSQL-only, inspired by other answers here.
SELECT all_ids AS missing_ids
FROM generate_series((SELECT MIN(id) FROM your_table), (SELECT MAX(id) FROM your_table)) all_ids
EXCEPT
SELECT id FROM your_table
I know it's an old question and already has an accepted answer,
but using a temp table isn't really necessary. Fixed formatting (sorry for double post).
DECLARE #TEST_ID integer, #LAST_ID integer, #ID integer
SET #TEST_ID = 1 -- start compare with this ID
SET #LAST_ID = 100 -- end compare with this ID
WHILE #TEST_ID <= #LAST_ID
BEGIN
SELECT #ID = (SELECT <column> FROM <table> WHERE <column> = #TEST_ID)
IF #ID IS NULL
BEGIN
PRINT 'Missing ID: ' + CAST(#TEST_ID AS VARCHAR(10))
END
SET #TEST_ID = #TEST_ID + 1
END
This is an Oracle only solution. It doesn't address the full question, but is left here for others that may be using Oracle.
select level id -- generate 1 .. 19
from dual
connect by level <= 19
minus -- remove from that set
select id -- everything that is currently in the
from table -- actual table
to get the missing rows from table
DECLARE #MaxID INT = (SELECT MAX(ID) FROM TABLE1)
SELECT SeqID AS MissingSeqID
FROM (SELECT ROW_NUMBER() OVER (ORDER BY column_id) SeqID from sys.columns) LkUp
LEFT JOIN dbo.TABLE1 t ON t.ID = LkUp.SeqID
WHERE t.ID is null and SeqID < #MaxID
I just have found the solution for Postgres:
select min(gs)
from generate_series(1, 1999) as gs
where gs not in (select id from mytable)
The single query can find the missing IDs..
SELECT distinct number
FROM master..spt_values
WHERE number BETWEEN 1 and (SELECT max(id) FROM MyTable)
AND number NOT IN (SELECT id FROM MyTable)
Update: This method took way too long so I wrote a linux command to find gaps in a text file. It does so in reverse order so first dump all id's to a text file like so;
nohup mysql --password=xx -e 'select id from tablename order by id desc' databasename > /home/ids.txt &
The first and last two lines are just to keep track of how long it took. 1.5million IDs(ish) took me 57sec & that's on a slow server. Set the max id in i and have at it.
T="$(date +%s)"; \
i=1574115; \
while read line; do \
if [[ "$line" != "$i" ]] ; then \
if [[ $i -lt 1 ]] ; then break; fi; \
if [[ $line -gt 1 ]] ; then \
missingsequenceend=$(( $line + 1 )); \
minusstr="-"; \
missingsequence="$missingsequenceend$minusstr$i"; \
expectnext=$(( $line - 1 )); \
i=$expectnext; \
echo -e "$missingsequence"; \
fi; \
else \
i=$(( $i - 1 )); \
fi; \
done \
< /home/ids.txt; \
T="$(($(date +%s)-T))"; \
echo "Time in seconds: ${T}"
Example output:
1494505-1494507
47566-47572
Time in seconds: 57
Also, I got syntax errors with the code from Eric's answer, but after changing the delimiter, using semicolons in the proper places and storing it in a procedure, it works.
Make sure you set the proper max ID, database name and table name (it's in the select query). And if you want to change the procedure name, change it in all 3 places.
use dbname;
drop procedure if exists dorepeat;
delimiter #
CREATE PROCEDURE dorepeat()
BEGIN
set #id = 1;
set #maxid = 1573736;
drop table if exists IDSeq;
create temporary table IDSeq
(
id int
);
WHILE #id < #maxid DO
insert into IDSeq values(#id);
set #id = #id + 1;
END WHILE;
select
s.id
from
IDSeq s
left join tablename t on
s.id = t.id
where t.id is null;
drop table if exists IDSeq;
END#
delimiter ;
CALL dorepeat;
I also found this query elwhere, but I haven't tested it.
SELECT a.id+1 AS start, MIN(b.id) - 1 AS end
FROM tablename AS a, tablename AS b
WHERE a.id < b.id
GROUP BY a.id
HAVING start < MIN(b.id)
TRY in MySQL
DELIMITER ||
DROP PROCEDURE IF EXISTS proc_missing ||
CREATE PROCEDURE proc_missing()
BEGIN
SET #minID = (SELECT MIN(`id`) FROM `tbl_name` WHERE `user_id`=13);
SET #maxID = (SELECT MAX(`id`) FROM `tbl_name` WHERE `user_id`=13);
REPEAT
SET #tableID = (SELECT `id` FROM `tbl_name` WHERE `id` = #minID);
IF (#tableID IS NULL) THEN
INSERT INTO temp_missing SET `missing_id` = #tableID;
END IF;
SET #minID = #minID + 1;
UNTIL(#minID <= #maxID)
END REPEAT;
END ||
DELIMITER ;
A few days ago, I was working on a production report and found some numbers missing. The missing numbers are very important, so I was asked to find a list of all missing numbers for investigation purposes. I posted a blog entry here, with a full demo, including a script to find missing numbers/IDs in a sample table.
The script suggested is quite long, so I won't include it here. Here are the basic steps used:
Create one temp table and store all distinct Numbers.
Find NextID which has something missing before it. Store into one TempTable.
Create one temp table to store missing number details.
Start to find the missing id using WHILE Loop.
Select missing data from #MissingID temp table.
Converting the SQL CTE (from Paul Svirin) to the Oracle version it looks like this (replace :YOURTABLE with the name of your table):
WITH Missing (missnum,maxid) as (
SELECT 1 missnum, (select max(id) from :YOURTABLE) maxid from dual
UNION ALL
SELECT m.missnum + 1,m.maxid
FROM Missing m
WHERE m.missnum < m.maxid
)
SELECT missnum
FROM Missing
LEFT OUTER JOIN :YOURTABLE tt on tt.id = Missing.missnum
WHERE tt.id is NULL
Using #PaulSvirin's answer, I've expanded it with a UNION to show ALL the data in my table, including the missing records with NULLs.
WITH Missing(missnum, maxid) AS
(SELECT (SELECT MIN(tmMIN.TETmeetingID)
FROM tblTETMeeting AS tmMIN)
AS missnum,
(SELECT MAX(tmMAX.TETmeetingID)
FROM tblTETMeeting AS tmMAX)
AS maxid
UNION ALL
SELECT missnum + 1, maxid
FROM Missing
WHERE missnum < maxid)
SELECT missnum AS TETmeetingID,
tt.DateID,
tt.WeekNo,
tt.TETID
FROM Missing LEFT JOIN tblTETMeeting tt ON tt.TETmeetingID = Missing.missnum
WHERE tt.TETmeetingID IS NULL
UNION
SELECT tt.TETmeetingID,
tt.DateID,
tt.WeekNo,
tt.TETID
FROM tblTETMeeting AS tt
OPTION ( MAXRECURSION 0 )
Work's great!
TETmeetingID DateID WeekNo TETID
29 3063 21 1
30 null null null
31 null null null
32 null null null
33 null null null
34 3070 22 1
35 3073 23 1
Easiest solution for me: Create a select that gives all ids up to max sequence value (ex:1000000), and filter:
with listids as (
Select Rownum idnumber From dual Connect By Rownum <= 1000000)
select * from listids
where idnumber not in (select id from table where id <=1000000)
A modified version borrowing #Eric proposal. This is for SQL Server and holds in a temp table the start and end value for missing ranges. If the gap is just one value it puts NULL as end value for easier visualization.
It will produce an output like this
|StartId| EndId |
|-------|-------|
| 1 | 10182 |
| 10189 | NULL |
| 10246 | 15000 |
And this is the script where myTable and id needs to be replaced by your table and identity column.
declare #id bigint
declare #endId bigint
declare #maxid bigint
declare #previousid bigint=0
set #id = 1
select #maxid = max(id) from myTable
create table #IDGaps
(
startId bigint,
endId bigint
)
while #id < #maxid
begin
if NOT EXISTS(select id from myTable where id=#id)
BEGIN
SET #previousid=#id
select top 1 #endId=id from myTable where id>#id
IF #id=#endId-1
insert into #IDGaps values(#id,null)
ELSE
insert into #IDGaps values(#id,#endId-1)
SET #id=#endId
END
ELSE
set #id = #id + 1
end
select * from #IDGaps
drop table #IDGaps
SOLUTION FOR SQLITE
if your table id only support positive values you can use this
SELECT DISTINCT table_id - 1 AS next_id
FROM table
WHERE next_id NOT IN (SELECT DISTINCT table_id FROM table)
AND next_id > 0
otherwise you should remove ids greater than the biggest id with
SELECT DISTINCT table_id + 1 AS next_id
FROM table
WHERE next_id NOT IN (SELECT DISTINCT table_id FROM table)
AND id < (SELECT MAX(id) FROM table)
This problem can be solved with only one query
select lft.id + 1 as missing_ids
from tbl as lft left outer join tbl as rght on lft.id + 1 = rght.id
where rght.id is null and lft.id between 1 and (Select max(id)-1 from tbl)
Tested on Mysql
Try This Query. This single query is enough to get missing numbers:(Please replace TABLE_NAME to which table name you are using)
select sno as missing from(SELECT #row := #row + 1 as sno FROM
(select 0 union all select 1 union all select 3 union all select 4 union all
select 5 union all select 6 union all select 6 union all select 7 union all
select 8 union all select 9) t,(select 0 union all select 1 union all select 3
union all select 4 union all select 5 union all select 6 union all select 6
union all select 7 union all select 8 union all select 9) t2,(select 0
union all select 1 union all select 3 union all select 4 union all select 5
union all select 6 union all select 6 union all select 7 union all select 8
union all select 9) t3, (select 0 union all select 1 union all select 3 union
all select 4 union all select 5 union all select 6 union all select 6 union all
select 7 union all select 8 union all select 9) t4,
(SELECT #row:=0) as b where #row<1000) as a where a.sno not in
(select distinct b.no from
(select b.*,if(#mn=0,#mn:=b.no,#mn) as min,(#mx:=b.no) as max from
(select ID as no from TABLE_NAME as a) as b,
(select #mn:=0,#mx:=0) as x order by no) as b) and
a.sno between #mn and #mx;
SELECT DISTINCT id -1
FROM users
WHERE id != 1 AND id - 1 NOT IN (SELECT DISTINCT id FROM users)
Explanation: ( id - 1 )..... checking for any previous id present in table
( id != 1 ).....neglecting when current id is 1 as its previous id will be 0 zero.
I have a large audit table and needed something that ran quickly - this worked well for me. It merges the top and bottom IDs for the missing ranges
select minQ.num,minId,maxId from
(SELECT DISTINCT id +1 as minId, Row_Number() Over ( Order By id ) As Num
FROM tblAuditLoghistory
WHERE id + 1 NOT IN (SELECT DISTINCT id FROM tblAuditLogHistory)
AND id < (SELECT max(id) FROM tblAuditLoghistory)) Minq
join
(SELECT DISTINCT id - 1 as maxId, Row_Number() Over ( Order By id ) As Num
FROM tblAuditLoghistory
WHERE id - 1 NOT IN (SELECT DISTINCT id FROM tblAuditLogHistory)
AND id > (SELECT min(id) FROM tblAuditLoghistory)) maxQ on minQ.num=maxQ.num
This what i used to find the missing id of one table named as tablename
select a.id+1 missing_ID from tablename a
where a.id+1 not in (select id from tablename b where b.id=a.id+1)
and a.id!=(select id from tablename c order by id desc limit 1)
It will return the missing ids.
If there are two(2) or more continuous missing ids, it will return only the first.