MySQL: Compare differences between two tables - sql

Question

Same as oracle diff: how to compare two tables? except in mysql.
Suppose I have two tables, t1 and t2 which are identical in layout but which may contain different data.
What's the best way to diff these two tables?
To be more precise, I'm trying to figure out a simple SQL query that tells me if data from one row in t1 is different from the data from the corresponding row in t2
It appears I cannot use the intersect nor minus. When I try
SELECT * FROM robot intersect SELECT * FROM tbd_robot
I get an error code:
[Error Code: 1064, SQL State: 42000] You have an error in your SQL
syntax; check the manual that corresponds to your MySQL server version
for the right syntax to use near 'SELECT * FROM tbd_robot' at line 1
Am I doing something syntactically wrong? If not, is there another query I can use?
Edit: Also, I'm querying through a free version DbVisualizer. Not sure if that might be a factor.

Answer 1

INTERSECT needs to be emulated in MySQL:
SELECT 'robot' AS `set`, r.*
FROM robot r
WHERE ROW(r.col1, r.col2, …) NOT IN
(
SELECT col1, col2, ...
FROM tbd_robot
)
UNION ALL
SELECT 'tbd_robot' AS `set`, t.*
FROM tbd_robot t
WHERE ROW(t.col1, t.col2, …) NOT IN
(
SELECT col1, col2, ...
FROM robot
)

Answer 2

You can construct the intersection manually using UNION. It's easy if you have some unique field in both tables, e.g. ID:
SELECT * FROM T1
WHERE ID NOT IN (SELECT ID FROM T2)
UNION
SELECT * FROM T2
WHERE ID NOT IN (SELECT ID FROM T1)
If you don't have a unique value, you can still expand the above code to check for all fields instead of just the ID, and use AND to connect them (e.g. ID NOT IN(...) AND OTHER_FIELD NOT IN(...) etc)

Answer 3

I found another solution in this link
SELECT MIN (tbl_name) AS tbl_name, PK, column_list
FROM
(
SELECT ' source_table ' as tbl_name, S.PK, S.column_list
FROM source_table AS S
UNION ALL
SELECT 'destination_table' as tbl_name, D.PK, D.column_list
FROM destination_table AS D
) AS alias_table
GROUP BY PK, column_list
HAVING COUNT(*) = 1
ORDER BY PK

Answer 4

select t1.user_id,t2.user_id
from t1 left join t2 ON t1.user_id = t2.user_id
and t1.username=t2.username
and t1.first_name=t2.first_name
and t1.last_name=t2.last_name
try this. This will compare your table and find all matching pairs, if any mismatch return NULL on left.

Answer 5

Based on Haim's answer I created a PHP code to test and display all the differences between two databases.
This will also display if a table is present in source or test databases.
You have to change with your details the <> variables content.
<?php
$User = "<DatabaseUser>";
$Pass = "<DatabasePassword>";
$SourceDB = "<SourceDatabase>";
$TestDB = "<DatabaseToTest>";
$link = new mysqli( "p:". "localhost", $User, $Pass, "" );
if ( mysqli_connect_error() ) {
die('Connect Error ('. mysqli_connect_errno() .') '. mysqli_connect_error());
}
mysqli_set_charset( $link, "utf8" );
mb_language( "uni" );
mb_internal_encoding( "UTF-8" );
$sQuery = 'SELECT TABLE_NAME FROM INFORMATION_SCHEMA.TABLES WHERE TABLE_SCHEMA="'. $SourceDB .'";';
$SourceDB_Content = query( $link, $sQuery );
if ( !is_array( $SourceDB_Content) ) {
echo "Table $SourceDB cannot be accessed";
exit(0);
}
$sQuery = 'SELECT TABLE_NAME FROM INFORMATION_SCHEMA.TABLES WHERE TABLE_SCHEMA="'. $TestDB .'";';
$TestDB_Content = query( $link, $sQuery );
if ( !is_array( $TestDB_Content) ) {
echo "Table $TestDB cannot be accessed";
exit(0);
}
$SourceDB_Tables = array();
foreach( $SourceDB_Content as $item ) {
$SourceDB_Tables[] = $item["TABLE_NAME"];
}
$TestDB_Tables = array();
foreach( $TestDB_Content as $item ) {
$TestDB_Tables[] = $item["TABLE_NAME"];
}
//var_dump( $SourceDB_Tables, $TestDB_Tables );
$LookupTables = array_merge( $SourceDB_Tables, $TestDB_Tables );
$NoOfDiscrepancies = 0;
echo "
<table border='1' width='100%'>
<tr>
<td>Table</td>
<td>Found in $SourceDB (". count( $SourceDB_Tables ) .")</td>
<td>Found in $TestDB (". count( $TestDB_Tables ) .")</td>
<td>Test result</td>
<tr>
";
foreach( $LookupTables as $table ) {
$FoundInSourceDB = in_array( $table, $SourceDB_Tables ) ? 1 : 0;
$FoundInTestDB = in_array( $table, $TestDB_Tables ) ? 1 : 0;
echo "
<tr>
<td>$table</td>
<td><input type='checkbox' ". ($FoundInSourceDB == 1 ? "checked" : "") ."></td>
<td><input type='checkbox' ". ($FoundInTestDB == 1 ? "checked" : "") ."></td>
<td>". compareTables( $SourceDB, $TestDB, $table ) ."</td>
</tr>
";
}
echo "
</table>
<br><br>
No of discrepancies found: $NoOfDiscrepancies
";
function query( $link, $q ) {
$result = mysqli_query( $link, $q );
$errors = mysqli_error($link);
if ( $errors > "" ) {
echo $errors;
exit(0);
}
if( $result == false ) return false;
else if ( $result === true ) return true;
else {
$rset = array();
while ( $row = mysqli_fetch_assoc( $result ) ) {
$rset[] = $row;
}
return $rset;
}
}
function compareTables( $source, $test, $table ) {
global $link;
global $NoOfDiscrepancies;
$sQuery = "
SELECT column_name,ordinal_position,data_type,column_type FROM
(
SELECT
column_name,ordinal_position,
data_type,column_type,COUNT(1) rowcount
FROM information_schema.columns
WHERE
(
(table_schema='$source' AND table_name='$table') OR
(table_schema='$test' AND table_name='$table')
)
AND table_name IN ('$table')
GROUP BY
column_name,ordinal_position,
data_type,column_type
HAVING COUNT(1)=1
) A;
";
$result = query( $link, $sQuery );
$data = "";
if( is_array( $result ) && count( $result ) > 0 ) {
$NoOfDiscrepancies++;
$data = "<table><tr><td>column_name</td><td>ordinal_position</td><td>data_type</td><td>column_type</td></tr>";
foreach( $result as $item ) {
$data .= "<tr><td>". $item["column_name"] ."</td><td>". $item["ordinal_position"] ."</td><td>". $item["data_type"] ."</td><td>". $item["column_type"] ."</td></tr>";
}
$data .= "</table>";
return $data;
}
else {
return "Checked but no discrepancies found!";
}
}
?>

Answer 6

Problem below, is to compare table before and after i do big update!.
If you use Linux, you can use commands as follow:
In terminal,
mysqldump -hlocalhost -uroot -p schema_name_here table_name_here > /home/ubuntu/database_dumps/dump_table_before_running_update.sql
mysqldump -hlocalhost -uroot -p schema_name_here table_name_here > /home/ubuntu/database_dumps/dump_table_after_running_update.sql
diff -uP /home/ubuntu/database_dumps/dump_some_table_after_running_update.sql /home/ubuntu/database_dumps/dump_table_before_running_update.sql > /home/ubuntu/database_dumps/diff.txt
You will need online tools for
Formatting SQL exported from the dumps,
e.g http://www.dpriver.com/pp/sqlformat.htm [Not the best I've seen]
We have diff.txt, you have to take manually the + - showing inside, which is 1 line of insert statements, that has the values.
Do diff online for the 2 lines - & + in diff.txt, past them in online diff tool
e.g https://www.diffchecker.com [you can save and share it, and has no limit on file size!]
Note: be extra careful if its sensitive/production data!

Answer 7

you can try The big data comparison platform in https://github.com/zhugezifang/dataCompare
this is a introduction of it
Design and practice of open source big data comparison platform
1. Background & current situation
In the process of developing large numbers, it is often encountered that data migration or upgrade, or different business parties have processed data according to their needs, but think that the data on both sides is still the same, so it will be necessary to manually compare the data. So is the data on both sides consistent? If not, what are the differences?
If there is no platform, you need to manually write some SQL scripts for comparison, and there is no evaluation standard. This is inefficient.
"Alibaba's Road to Big Data" actually mentions such a platform, but because it is not used externally, the introduction in the book is relatively simple. Based on previous work experience, a big data comparison platform was developed to assist in verifying data, named dataCompare.
Main solutions:
(1) Verify data and data comparison, which wastes great labor costs
(2) Without a set of standards, the results of verification are difficult to evaluate
(3) Automatic data verification and comparison can be achieved by interface interaction, check or low-code
[enter image description here][1]
2. Purpose
(1) Automatic data verification and comparison can be achieved by interface interaction, check or low-code.
(2) The data team's data comparison efficiency is increased by at least about 50%.
(3) A set of unified data verification scheme to meet the standard specifications of data verification and comparison
3. System architecture design
4. The current version has implemented the following functions
(1) Low-code simple configuration completes the core function of data comparison
(2) Data magnitude comparison and data consistency comparison
5. Follow-up Development Plan
(1) Discrepancy case finding
(2) Data pointer detection---- enumeration value detection, range detection, numerical detection, primary key mode detection
(3) Data comparison task is scheduled and automatically scheduled
(4) Automatically send an email report to the comparison results
6. The core code is opening in githup
https://github.com/zhugezifang/dataCompare
[enter image description here][1]

Answer 8

Based on Haim's answer here's a simplified example if you're looking to compare values that exist in BOTH tables, otherwise if there's a row in one table but not the other it will also return it....
Took me a couple of hours to figure out. Here's a fully tested simply query for comparing "tbl_a" and "tbl_b"
SELECT ID, col
FROM
(
SELECT
tbl_a.ID, tbl_a.col FROM tbl_a
UNION ALL
SELECT
tbl_b.ID, tbl_b.col FROM tbl_b
) t
WHERE ID IN (select ID from tbl_a) AND ID IN (select ID from tbl_b)
GROUP BY
ID, col
HAVING COUNT(*) = 1
ORDER BY ID
So you need to add the extra "where in" clause:
WHERE ID IN (select ID from tbl_a) AND ID IN (select ID from tbl_b)
Also:
For ease of reading if you want to indicate the table names you can use the following:
SELECT tbl, ID, col
FROM
(
SELECT
tbl_a.ID, tbl_a.col, "name_to_display1" as "tbl" FROM tbl_a
UNION ALL
SELECT
tbl_b.ID, tbl_b.col, "name_to_display2" as "tbl" FROM tbl_b
) t
WHERE ID IN (select ID from tbl_a) AND ID IN (select ID from tbl_b)
GROUP BY
ID, col
HAVING COUNT(*) = 1
ORDER BY ID

Answer 9

you can user my own developed tool
https://github.com/hardeepvicky/MySql-Schema-Compare

Answer 10

I tried the above answer but found that if one table has null values and the second table has values in a column then the intersect code above does not report this fact.
select p.pcn,p.period,p.account_no,p.ytd_debit,a.ytd_debit
-- select count(*) -- 157,283
from Plex.account_period_balance p -- 157,283/202207,148,998
join Azure.account_period_balance a -- 157,283/202207,148,998
on p.pcn = a.pcn
and p.period = a.period
and p.account_no = a.account_no -- 157,283
where p.period_display = a.period_display -- 157,283
and p.debit = a.debit -- 157,283
-- and p.ytd_debit = a.ytd_debit -- 148,998
-- and p.ytd_debit != a.ytd_debit -- 0