I have a Dynamic Pivot in MySQL (see this question: MySQL "Multi-Dimensional?" Dynamic Pivot)
I want to know the Column/As/Field names as if it were a table and I queried INFORMATION_SCHEMA (which if this was a REAL table, would be easy enough: MySQL query to get column names?).
But, I can find no question or reference to a function or SELECT option to get the Column/As/Field names from a query. Is this possible? If so, how?
Using Perl to access MySQL (http://dbi.perl.org/).
So, flipping this around... we know the fixed columns. So, if we use the same basic query that creates the Pivot to begin with, we can get a GROUP_CONCAT of the names.
SET #sql = NULL;
SELECT GROUP_CONCAT(qrv.req_name) INTO #sql
FROM (SELECT qrt.req_name FROM qual_requirment_values qrv JOIN qual_requirment_types qrt ON qrt.id = qrv.req_type_id) qrv;
SET #sql = CONCAT('r.rank,r.member_type,im.name,qrv.grouping,', #sql);
SELECT #sql;
This can then be split into an array and used.
Seems like the long way around, but in the absence of something else it will work for my application. Seems horribly inefficient! :)
The better answer, thanks to #OllieJones. The Data Base Interface used to access MySQL should provide a way.
In my case (Perl), the answer is here: http://www.perlmonks.org/?node_id=264623
my $sql = ... [some query];
my $sth = $dbh->prepare($sql);
$sth->execute();
my $field_name_arrayref = $sth->{NAME};
Further to the answer, this is the full method within my MySQL package. do() is our generic DBI method that returns queries in an AoA. Adapting that method to create do_fieldNames();
## Tested Method
sub do_fieldNames {
my ($self, $sql, $has_results) = #_;
my ($sth, $rv, #row, #query_results);
## Execute the SQL statement
$sth = $$self->prepare($sql);
$rv = $sth->execute or &error(3306, __LINE__, __FILE__, $sql, $$self->errstr);
return undef unless $rv > 0;
## SOLUTION >> Field Name arrayref, part of the standard included DBI Perl Module
my $field_name_arrayref = $sth->{NAME};
## Parse the results
if ($has_results || $sql =~ /^select/i) {
while (#row = $sth->fetchrow_array) {
push #query_results, [ #row ];
}
}
## Return results
return (\#query_results, $field_name_arrayref) ;
}
Using MySQL I can run the query:
SHOW CREATE TABLE MyTable;
And it will return the create table statement for the specificed table. This is useful if you have a table already created, and want to create the same table on another database.
Is it possible to get the insert statement for an already existing row, or set of rows? Some tables have many columns, and it would be nice for me to be able to get an insert statement to transfer rows over to another database without having to write out the insert statement, or without exporting the data to CSV and then importing the same data into the other database.
Just to clarify, what I want is something that would work as follows:
SHOW INSERT Select * FROM MyTable WHERE ID = 10;
And have the following returned for me:
INSERT INTO MyTable(ID,Col1,Col2,Col3) VALUES (10,'hello world','some value','2010-10-20');
There doesn't seem to be a way to get the INSERT statements from the MySQL console, but you can get them using mysqldump like Rob suggested. Specify -t to omit table creation.
mysqldump -t -u MyUserName -pMyPassword MyDatabase MyTable --where="ID = 10"
In MySQL Workbench you can export the results of any single-table query as a list of INSERT statements. Just run the query, and then:
click on the floppy disk near Export/Import above the results
give the target file a name
at the bottom of the window, for Format select SQL INSERT statements
click Save
click Export
Since you copied the table with the SQL produced by SHOW CREATE TABLE MyTable, you could just do the following to load the data into the new table.
INSERT INTO dest_db.dest_table SELECT * FROM source_db.source_table;
If you really want the INSERT statements, then the only way that I know of is to use mysqldump http://dev.mysql.com/doc/refman/5.1/en/mysqldump.htm. You can give it options to just dump data for a specific table and even limit rows.
I wrote a php function that will do this. I needed to make an insert statement in case a record needs to be replaced after deletion for a history table:
function makeRecoverySQL($table, $id)
{
// get the record
$selectSQL = "SELECT * FROM `" . $table . "` WHERE `id` = " . $id . ';';
$result = mysql_query($selectSQL, $YourDbHandle);
$row = mysql_fetch_assoc($result);
$insertSQL = "INSERT INTO `" . $table . "` SET ";
foreach ($row as $field => $value) {
$insertSQL .= " `" . $field . "` = '" . $value . "', ";
}
$insertSQL = trim($insertSQL, ", ");
return $insertSQL;
}
Laptop Lift's code works fine, but there were a few things I figured people may like.
Database handler is an argument, not hardcoded. Used the new mysql api. Replaced $id with an optional $where argument for flexibility. Used real_escape_string in case anyone has ever tried to do sql injection and to avoid simple breakages involving quotes. Used the INSERT table (field...) VALUES (value...)... syntax so that the fields are defined only once and then just list off the values of each row (implode is awesome). Because Nigel Johnson pointed it out, I added NULL handling.
I used $array[$key] because I was worried it might somehow change, but unless something is horribly wrong, it shouldn't anyway.
<?php
function show_inserts($mysqli,$table, $where=null) {
$sql="SELECT * FROM `{$table}`".(is_null($where) ? "" : " WHERE ".$where).";";
$result=$mysqli->query($sql);
$fields=array();
foreach ($result->fetch_fields() as $key=>$value) {
$fields[$key]="`{$value->name}`";
}
$values=array();
while ($row=$result->fetch_row()) {
$temp=array();
foreach ($row as $key=>$value) {
$temp[$key]=($value===null ? 'NULL' : "'".$mysqli->real_escape_string($value)."'");
}
$values[]="(".implode(",",$temp).")";
}
$num=$result->num_rows;
return "INSERT `{$table}` (".implode(",",$fields).") VALUES \n".implode(",\n",$values).";";
}
?>
In PHPMyAdmin you can:
click copy on the row you want to know its insert statements SQL:
click Preview SQL:
you will get the created insert statement that generates it
You can apply that on many rows at once if you select them and click copy from the bottom of the table and then Preview SQl
I use the program SQLYOG where I can make a select query, point at the results and choose export as sql. This gives me the insert statements.
The below command will dump into the terminal without all the extra stuff mysqldump will output surrounding the INSERT. This allows copying from the terminal without it writing to a file. This is useful if the environment restricts writing new files.
mysqldump -u MyUserName -pMyPassword MyDatabase MyTable --where="ID = 10" --compact --no-create-info --complete-insert --quick
Using mysqldump --help I found the following options.
-q, --quick Don't buffer query, dump directly to stdout.
(Defaults to on; use --skip-quick to disable.)
-t, --no-create-info
Don't write table creation info.
-c, --complete-insert
Use complete insert statements.
--compact Give less verbose output (useful for debugging). Disables
structure comments and header/footer constructs. Enables
options --skip-add-drop-table --skip-add-locks
--skip-comments --skip-disable-keys --skip-set-charset.
If you want get "insert statement" for your table you can try the following code.
SELECT
CONCAT(
GROUP_CONCAT(
CONCAT(
'INSERT INTO `your_table` (`field_1`, `field_2`, `...`, `field_n`) VALUES ("',
`field_1`,
'", "',
`field_2`,
'", "',
`...`,
'", "',
`field_n`,
'")'
) SEPARATOR ';\n'
), ';'
) as `QUERY`
FROM `your_table`;
As a result, you will have insers statement:
INSERT INTO your_table (field_1, field_2, ..., field_n) VALUES (value_11, value_12, ... , value_1n);
INSERT INTO your_table (field_1, field_2, ..., field_n) VALUES (value_21, value_22, ... , value_2n);
/...................................................../
INSERT INTO your_table (field_1, field_2, ..., field_n) VALUES (value_m1, value_m2, ... , value_mn);
, where m - number of records in your_table
Within MySQL work bench perform the following:
Click Server > Data Export
In the Object Selection Tab select the desired schema.
Next, select the desired tables using the list box to the right of the schema.
Select a file location to export the script.
Click Finish.
Navigate to the newly created file and copy the insert statements.
you can use Sequel pro to do this, there is an option to 'get as insert statement' for the results obtained
There is a quite easy and useful solution for creating an INSERT Statement for editing without the need to export SQL with just Copy & Paste (Clipboard):
Select the row in a query result window of MySQL Workbench, probably even several rows. I use this even if the row does contain different data than I want to insert in my script or when the goal is to create a prepared statement with ? placeholders.
Paste the copied row which is in your clipboard now into the same table (query results list is an editor in workbench) into the last free row.
Press "Apply" and a windows opens showing you the INSERT statement - DO NOT EXECUTE
Copy the SQL from the window to your clipboard
CANCEL the execution, thus not changing the database but keeping the SQL in your clipboard.
Paste the SQL wherever you want and edit it as you like, like e.g. inserting ? placeholders
You can create a SP with the code below - it supports NULLS as well.
select 'my_table_name' into #tableName;
/*find column names*/
select GROUP_CONCAT(column_name SEPARATOR ', ') from information_schema.COLUMNS
where table_schema =DATABASE()
and table_name = #tableName
group by table_name
into #columns
;
/*wrap with IFNULL*/
select replace(#columns,',',',IFNULL(') into #selectColumns;
select replace(#selectColumns,',IFNULL(',',\'~NULL~\'),IFNULL(') into #selectColumns;
select concat('IFNULL(',#selectColumns,',\'~NULL~\')') into #selectColumns;
/*RETRIEVE COLUMN DATA FIELDS BY PK*/
SELECT
CONCAT(
'SELECT CONCAT_WS(','''\'\',\'\''',' ,
#selectColumns,
') AS all_columns FROM ',#tableName, ' where id = 5 into #values;'
)
INTO #sql;
PREPARE stmt FROM #sql;
EXECUTE stmt;
/*Create Insert Statement*/
select CONCAT('insert into ',#tableName,' (' , #columns ,') values (\'',#values,'\')') into #prepared;
/*UNWRAP NULLS*/
select replace(#prepared,'\'~NULL~\'','NULL') as statement;
For HeidiSQL users:
If you use HeidiSQL, you can select the row(s) you wish to get insert statement. Then right click > Export grid rows > select "Copy to clipboard" for "Output target", "Selection" for "Row Selection" so you don't export other rows, "SQL INSERTs" for "Output format" > Click OK.
The insert statement will be inside you clipboard.
In case you use phpMyAdmin (Tested on version 5.x):
Click on "Edit" button next to the row for which you would like to have an insert statement, then on the bottom next to the action buttons just select "Show insert query" and press "Go".
With PDO you can do it this way.
$stmt = DB::getDB()->query("SELECT * FROM sometable", array());
$array = $stmt->fetchAll(PDO::FETCH_ASSOC);
$fields = array_keys($array[0]);
$statement = "INSERT INTO user_profiles_copy (".implode(",",$fields).") VALUES ";
$statement_values = null;
foreach ($array as $key => $post) {
if(isset($statement_values)) {
$statement_values .= ", \n";
}
$values = array_values($post);
foreach($values as $index => $value) {
$quoted = str_replace("'","\'",str_replace('"','\"', $value));
$values[$index] = (!isset($value) ? 'NULL' : "'" . $quoted."'") ;
}
$statement_values .= "(".implode(',',$values).")";
}
$statement .= $statement_values . ";";
echo $statement;
I think that the answer provided by Laptop Lifts is best...but since nobody suggested the approach that I use, i figured i should chime in. I use phpMyAdmin to set up and manage my databases most of the time. In it, you can simply put checkmarks next to the rows you want, and at the bottom click "Export" and chose SQL. It will give you INSERT statements for whichever records you selected. Hope this helps.
You can try this
function get_insert_query($pdo, $table, $where_sth)
{
$sql = "";
$row_data = $pdo->query("SELECT * FROM `{$table}` WHERE $where_sth")->fetch();
if($row_data){
$sql = "INSERT INTO `$table` (";
foreach($row_data as $col_name => $value){
$sql .= "`".$col_name."`, ";
}
$sql = rtrim($sql, ", ");
$sql .= ") VALUES (";
foreach($row_data as $col_name => $value){
if (is_string($value)){
$value = $pdo->quote($value);
} else if ($value === null){
$value = 'NULL';
}
$sql .= $value .", ";
}
$sql = rtrim($sql, ", ");
$sql .= ");";
}
return $sql;
}
To use it, just call:
$pdo = new PDO( "connection string goes here" );
$sql = get_insert_query($pdo, 'texts', "text_id = 959");
echo $sql;
Update for get insert statement for current registers at PhpMyAdmin:
Select the table from you DB to get registers from
"Export" tab at the top menú as the image below
Custom export
Move down to "Data creation options"
Once there, select "Insert" function at your preferred syntax
Its very simple. All you have to do is write an Insert statement in a static block and run it as a script as a whole block.
E.g. If you want to get Insert statements from a table (say ENGINEER_DETAILS) for selected rows, then you have to run this block -
spool
set sqlformat insert
select * from ENGINEER_DETAILS where engineer_name like '%John%';
spool off;
The output of this block will be Insert statements.
In MySQL Workbench, right-click the table and select 'Send to SQL Editor'/'Insert Statement'. Clean it up a bit and you're good to go.
Based on your comments, your goal is to migrate database changes from a development environment to a production environment.
The best way to do this is to keep your database changes in your source code and consequently track them in your source control system such as git or svn.
you can get up and running quickly with something like this: https://github.com/davejkiger/mysql-php-migrations
as a very basic custom solution in PHP, you can use a function like this:
function store_once($table, $unique_fields, $other_fields=array()) {
$where = "";
$values = array();
foreach ($unique_fields as $k => $v) {
if (!empty($where)) $where .= " && ";
$where .= "$k=?";
$values[] = $v;
}
$records = query("SELECT * FROM $table WHERE $where", $values);
if (false == $records) {
store($table, array_merge($unique_fields, $other_fields));
}
}
then you can create a migration script which will update any environment to your specifications.
What is the preferred way to insert strings that can contain both single and double quotes (",') into MySql using DBI? For example, $val1 and $val2 can contain quotes:
my $dbh = DBI->connect( ... );
my $sql = "insert into tbl_name(col_one,col_two) values($val1, $val2)";
my $sth = $dbh->prepare($sql);
$sth->execute();
Use a bound query using
$sth = $dbh->prepare("insert into tbl_name(col_one,col_two) values(?,?)");
$sth->execute($val1, $val2);
If you use bound variables, everything is escaped for you.
Update: Changed my example to correspond with the example edited into the question.
Update: I don't know why Adam deleted his answer, but if for some reason you can't use bound variables (aka "placeholders"), you can also use $dbh->quote($var) on the variable. For example:
$sql = sprintf "SELECT foo FROM bar WHERE baz = %s",
$dbh->quote(q("Don't"));
Use the quote() method. It will intelligently handle the quoting for you. Example from the docs:
$sql = sprintf "SELECT foo FROM bar WHERE baz = %s",
$dbh->quote("Don't");
Slightly modified to have both types of quotes:
$sql = sprintf "SELECT foo FROM bar WHERE baz = %s",
$dbh->quote(q("Don't"));
One small caveat on the bound placeholders, I build a rather large database-loading script that initially used bound placeholders in an older version of Perl/DBI and found what appears to be a memory leak in the placeholder implementation, so if you're looking at using them in a persistent process/daemon or in a high-volume context you may want to make sure process size doesn't become an issue. Switching over to building the query strings using the quote() method eliminated the issue for me.
DBI placeholders are awesome. They shine when you need to execute the same query in a loop. Consider this:
my $dbh = DBI->connect(...);
my $name_pairs = get_csv_data("data.csv");
my $sth = $dbh->prepare("INSERT INTO t1 (first_name, last_name) VALUES (?,?)");
for my $pair (#$name_pairs) {
unless ($sth->execute(#$pair)) {
warn($sth->errstr);
}
}
In this case, having the prepared statement handle is, er, handy.
However, barring this sort of tight-loop cases, I like to see the actual statement that was sent to the server. This is where I lean heavily on quote and frankly sprintf.
# Here, I am confident about the hash keys, less so about the values
$sql = sprintf("INSERT INTO t1 (%s) VALUES (%s)",
join(",", keys(%hash)),
join("," map { $dbh->quote($_) } values(%hash))
);
$sth = $dbh->prepare($sql);
unless ($sth->execute) {
warn($sth->{Statement});
}
Note that you do have to set RaiseError => 0 on $dbh so that you can see the SQL that failed, but this has helped me a great deal in the past.
Cheers.