For example, if I need to join TableA & TableB - how can I use the INFORMATION_SCHEMA database to find the common field between them?
Here is one method:
select column_name
from information_schema.columns c
where table_name in ('A', 'B')
group by column_name
having count(*) = 2;
If necessary, you should also include table_schema to identify the tables.
Another way to get common fields in two table:
SELECT C.name
FROM SYS.OBJECTS O
JOIN SYS.COLUMNS C
ON O.object_id = C.object_id
WHERE O.name IN
(
'TABLE-A', 'TABLE-B'
)
group by C.name
having count(*) = 2;
Way to get common column in two table:
SELECT A.COLUMN_NAME
FROM INFORMATION_SCHEMA.COLUMNS A
JOIN INFORMATION_SCHEMA.COLUMNS B ON A.COLUMN_NAME = B.COLUMN_NAME
WHERE A.TABLE_NAME 'Table_1'
AND B.TABLE_NAME = 'Table_2'
Related
There are 2 tables which have 10 columns common in both the tables and 2 columns exist in table 2 but not in table 1. I have a task where I need to get those 2 columns present in Table 2 but not in Table 1.
The condition for the query is:
We must use information_schema.columns table.
Left join must be used on the same table information_schema.columns
SELECT *
FROM INFORMATION_SCHEMA.COLUMNS t2
LEFT JOIN INFORMATION_SCHEMA.COLUMNS t1
ON t2.COLUMN_NAME = t1.COLUMN_NAME
AND t1.TABLE_NAME = 'table1'
WHERE t2.TABLE_NAME = 'table2'
AND t1.COLUMN_NAME IS NULL;
As per your requirement there care multiple solutions to this problem. Just giving a simple one:
Left join the information_schema.columns table with itself on the table_name and column_names and then filter out the columns of 'table1' using where not exists from the joined result. Finally apply another filter on the joined result to take only the column_name which are in 'table2':
select isc1.column_name from information_schema.columns isc1
left join information_schema.columns isc2
on isc1.table_name = isc2.table_name and isc1.column_name = isc2.column_name
where not exists (select column_name
from information_schema.columns isc
where isc.table_name = 'table1')
and isc1.table_name = 'table2';
A simple way is to take all the columns of Table 2 that are not in Table 1.
SELECT c.COLUMN_NAME
FROM information_schema.tables t
join INFORMATION_SCHEMA.COLUMNS c on t.TABLE_NAME = c.TABLE_NAME
WHERE t.TABLE_NAME = 'table2' and c.COLUMN_NAME not in
(
SELECT c.COLUMN_NAME
FROM information_schema.tables t
join INFORMATION_SCHEMA.COLUMNS c on t.TABLE_NAME = c.TABLE_NAME
WHERE t.TABLE_NAME = 'table1'
)
Need to also select the row/record count using the following query in Oracle SQL. How would I go about doing this?
SELECT DISTINCT t.OWNER AS TABLE_SCHEMA, t.Table_Name, c.COLUMN_NAME, c.DATA_TYPE,
(
SELECT CASE WHEN cons.CONSTRAINT_TYPE = 'P' THEN 'Primary Key' ELSE NULL END
FROM SYS.all_cons_columns cols
INNER JOIN SYS.all_constraints cons ON cons.constraint_name=cols.constraint_name AND cons.constraint_type = 'P'
WHERE cols.Column_name=c.column_name AND cols.TABLE_NAME=c.TABLE_NAME AND cols.OWNER=c.OWNER
) AS CONSTRAINT_TYPE, c.DATA_PRECISION, c.DATA_SCALE
FROM SYS.ALL_TABLES t
INNER JOIN SYS.all_tab_columns c ON c.TABLE_NAME=t.TABLE_NAME AND c.OWNER=t.OWNER
WHERE t.OWNER = 'MY_SCHEMA_NAME' AND t.Table_Name = 'MY_TABLE_NAME'
If you can live with an estimate number of rows (whose accuracy depends on the freshness of each table's statistics), you can just use column NUM_ROWS from ALL_TABLES:
SELECT
t.OWNER AS TABLE_SCHEMA,
t.Table_Name,
c.COLUMN_NAME,
c.DATA_TYPE,
t.NUM_ROWS, --> here
...
FROM all_tables t
INNER JOIN all_tab_columns c ...
Note that DISTINCT seem superfluous here - your query generates one record per column in the table.
If you really need the exact count, then you can use an inline query - but you need to hardcode the table name:
SELECT
t.OWNER AS TABLE_SCHEMA,
t.Table_Name,
c.COLUMN_NAME,
c.DATA_TYPE,
(SELECT COUNT(*) FROM my_schema_name.my_table_name) no_rows,
...
FROM all_tables t
INNER JOIN all_tab_columns c ...
WHERE t.OWNER = 'MY_SCHEMA_NAME' AND t.Table_Name = 'MY_TABLE_NAME'
Not sure which SQL you are using but you might put the whole result into a temp table
like
INTO temp temp_table_count WITH NO LOG;
SELECT COUNT(*) FROM temp_table_count;
https://www.oracletutorial.com/oracle-basics/oracle-private-temporary-table/ the reference
INSERT INTO ora$ppt_temp_table_count(ADD ALL COLUMNS HERE
COMMASEPARATED)
SELECT col1, col2, col3
FROM source_table
WHERE condition; <-- your query
SELECT COUNT(*) FROM ora$ppt_temp_table_count;
I am using SQL Server 2017. How can I search for multiple column names in a single table?
I can search for a single column name (USERID) in my database with:
select *
from INFORMATION_SCHEMA.COLUMNS
where COLUMN_NAME like '%USERID%'
order by TABLE_NAME
But, I want to search for all tables that have both USERID and DOB.
Using:
select *
from INFORMATION_SCHEMA.COLUMNS
where COLUMN_NAME like '%USERID%' OR COLUMN_NAME like '%DOB%'
order by TABLE_NAME
returns a list of all tables that contain the column USERID as well as all tables that contain DOB.
Changing to AND results in a completely empty list, which is correct, as a single column would not contain both terms
You can try below -
select TABLE_NAME from INFORMATION_SCHEMA.COLUMNS
where COLUMN_NAME like '%USERID%' OR COLUMN_NAME like '%DOB%'
group by TABLE_NAME
having count(distinct COLUMN_NAME)=2
Below query will give you desired results:
;WITH cte
AS (SELECT object_id , ROW_NUMBER() OVER(PARTITION BY object_id
ORDER BY object_id) AS rn
FROM sys.columns AS c
WHERE c.name = 'USERID'
OR
c.name = 'DOB')
SELECT s.name + '.' + t.name
FROM sys.tables AS t INNER JOIN cte AS c ON t.object_id = c.object_id
INNER JOIN sys.schemas AS s ON t.schema_id = s.schema_id
WHERE c.rn > 1;
Please note to use exact column names with = operator and not like operator to get the correct results.
In my DB test, I have 400 tables.I would like to know in what table the column ID exists and in which tables the column ID does not exist. (see the pic)
I tryed with this query, but doesn't work. What's the best way ?
SELECT COLUMNS.TABLE_NAME, 'Yes' As MyCol_ID
FROM INFORMATION_SCHEMA.COLUMNS COLUMNS,
INFORMATION_SCHEMA.TABLES TABLES
WHERE COLUMNS.TABLE_NAME = TABLES.TABLE_NAME
AND Upper(COLUMNS.COLUMN_NAME) = Upper('ID')
AND TABLES.TABLE_TYPE='BASE TABLE'
order by COLUMNS.TABLE_NAME
UNION
SELECT COLUMNS.TABLE_NAME, 'No' As MyCol_ID
FROM INFORMATION_SCHEMA.COLUMNS COLUMNS,
INFORMATION_SCHEMA.TABLES TABLES
WHERE COLUMNS.TABLE_NAME = TABLES.TABLE_NAME
AND Upper(COLUMNS.COLUMN_NAME) <> Upper('ID')
AND TABLES.TABLE_TYPE='BASE TABLE'
order by COLUMNS.TABLE_NAME
One method is basically a left join:
select t.table_schema, t.table_name,
(case when c.column_name is null then 'no' else 'yes' end) as id_exists
from information_schema.tables t left join
information_schema.columns c
on t.table_name = c.table_name and t.table_schema = c.table_schema and
c.column_name = 'ID'
where t.table_type = 'BASE_TABLE';
Notice that I fixed the join to include the schema name as well as the table name.
I like using the sys. variants. Run this in the context of your test DB:
SELECT T.NAME, CASE WHEN C.NAME = 'my_ColID' THEN 'YES' ELSE 'NO' END
FROM SYS.TABLES T
LEFT JOIN SYS.COLUMNS C ON C.object_id = T.object_id
AND C.NAME = 'my_ColID'
ORDER BY T.NAME
I am creating a query that returns the number of columns in each table, but I want to exclude Views.
The following works but returns View results:
SELECT COUNT(*), table_name
FROM INFORMATION_SCHEMA.COLUMNS
Group By table_name
Any suggestions?
NOTE: MSSQL 2005+
This assumes SQL 2005 or higher
SELECT
t.name,
count(c.name)
FROM
sys.tables t
inner join sys.columns c
ON t.object_id = c.object_id
group by t.name
Something like this:
SELECT COUNT(col.column_name), col.table_name
FROM information_schema.columns col
JOIN information_schema.tables tbl
ON tbl.table_name = col.table_name
AND tbl.table_schema = col.table_schema
AND tbl.table_catalog = col.table_catalog
AND tbl.table_type <> 'VIEW'
GROUP BY col.table_name
Join INFORMATION_SCHEMA.TABLES to find out if the table is a view.
SELECT COUNT(col.column_name), tab.table_name
FROM INFORMATION_SCHEMA.tables tab
JOIN INFORMATION_SCHEMA.COLUMNS col ON col.table_name = tab.table_name
WHERE tab.table_type != 'VIEW'
GROUP BY 2
ORDER BY 2
SELECT tab.table_name,COUNT(col.column_name)
FROM INFORMATION_SCHEMA.tables tab
JOIN INFORMATION_SCHEMA.COLUMNS col ON col.table_name = tab.table_name
WHERE tab.table_type != 'VIEW'
GROUP BY tab.table_name
ORDER BY tab.table_name
This code shows the table names and their columns count in front of them.
you can add the schema name if you want.