SQLite: Add columns on the fly if they don't exist - perl

sub insert {
my ($self) = #_;
my $data = $self->{data};
my $keys = join( ', ', keys $data);
my $values = join( ', ', map qq('$_'), values $data);
my $sql = "INSERT INTO tbl ($keys) VALUES ($values);";
my $sth = $self->{dbh}->prepare($sql);
$sth->execute();
}
I have a method that inserts the contents of a hash ref into a one table sqlite database. I was wondering if there was a simple way to add a column to the table if the hash key is not already a column. Obviously if one of the keys is not a column name, the insert will fail. Can I capitalize on that failure, add the missing columns, and redo the insert. Or would I have to check all columns against all the keys each time I want to insert into the database? (All keys have TEXT values)

Alter your table based on the info you get using PRAGMA
my $inf_query = $db->prepare("PRAGMA table_info('tbl')");
$inf_query->execute();
my #inf = map { $_->[1] } #{$inf_query->fetchall_arrayref()};
#inf will be an array containing the columns present in the table, and you can use that info to construct your ALTER query.
Edited to return an array you can use to grep ;)

Related

Perl DBI and temporary tables

I am using the Perl DBI and DB2.
When I run this code:
sub MergePolygonNameTable()
{
my $table = "THESCHEMA.NAME";
print "Merging into ${table} table. ", scalar localtime, "\n";
eval
{
$DbHandle->do("
declare global temporary table session.TEMP_NAME
(POLICY_MASTER_ID INT
)
;
");
$DbHandle->do("
CREATE UNIQUE INDEX session.TEMP_NAME_IDX1 ON session.TEMP_NAME
(POLICY_MASTER_ID ASC
)");
$DbHandle->do("
insert into session.TEMP_NAME
(POLICY_MASTER_ID
)
SELECT pm.ID as POLICY_MASTER_ID
FROM THESCHEMA.POLICY_MASTER pm
");
$DbHandle->do("
MERGE INTO THESCHEMA.NAME as t
USING session.TEMP_NAME as s
ON t.POLICY_MASTER_ID = s.POLICY_MASTER_ID
WHEN MATCHED
) THEN
UPDATE SET t.UPDATED_DATETIME = CURRENT_TIMESTAMP
WHEN NOT MATCHED THEN
INSERT
(POLICY_MASTER_ID
) VALUES
(s.POLICY_MASTER_ID
)
;
");
};
if ($#)
{
print STDERR "ERROR: $ExeName: Cannot merge into ${table} table.\n$#\n";
ExitProc(1);
}
}
The problem is that the THESCHEMA.NAME is empty after the run.
I suspect that DBI does not keep the contents of the temporary table after the do(). But the DBI does not allow me to put more than one statement in a do().
How do I get temporary tables to work in the DBI?
The ON COMMIT DELETE ROWS is the default option of the DECLARE GLOBAL TEMPORARY TABLE statement.
Use the ON COMMIT PRESERVE ROWS option to preserve the rows upon explicit or implicit COMMIT.
Like this:
declare global temporary table session.TEMP_NAME
(POLICY_MASTER_ID INT
)
ON COMMIT PRESERVE ROWS
;

Add char to the array elements

Is there any easier way to add char to the beginning and end of array elements in perl.
Ex:my $str ='table1,table2,table3'
my #arry = split (/,/,$str)
I like the output should look like
'table1','table2','table3'
So it can be used in sql query
Thanks.
join(',', map "'$_'", #arry)
is what you are asking for.
But it's much better to use placeholders:
my $str = 'table1,table2,table3';
my #arry = split(/,/, $str);
if (#arry) {
my $query = 'select * from tablename where colname in (' . join(',',('?') x #arry) . ')';
my $sth = $dbh->prepare($query);
$sth->execute(#arry);
...
}
If you are going to use DBI (https://metacpan.org/module/DBI) to work with your database, it can do it for you if you use placeholders in your queries, so that you don't have to quote values.
For example, if you want to insert something into a table:
$dbh->do("INSERT INTO table VALUES(?, ?, ?)", undef, #arry)
or die $dbh->errstr;
where #arry is an array containing exactly 3 values to be inserted into a table.
More on this can be found here https://metacpan.org/module/DBI#Placeholders-and-Bind-Values

DBI: selectall_arrayref and columnnames

When I fetch the data this way is it possible then to access the column names and the column types or do I need an explicit prepare to reach this?
use DBI;
my $dbh = DBI->connect( ... );
my $select = "...";
my #arguments = ( ... );
my $ref = $dbh->selectall_arrayref( $select, {}, #arguments, );
Update:
With prepare I would do it this way:
my $sth = $dbh->prepare( $select );
$sth->execute( #arguments );
my $col_names = $sth->{NAME};
my $col_types = $sth->{TYPE};
my $ref = $sth->fetchall_arrayref;
unshift #$ref, $col_names;
The best solution is to use prepare to get a statement handle, as you describe in the second part of your question. If you use selectall_hashref or selectall_arrayref, you don't get a statement handle, and have to query the column type information yourself via $dbh->column_info (docs):
my $sth = $dbh->column_info('','',$table,$column); # or $column='' for all
my $info = $sth->fetchall_arrayref({});
use Data::Dumper; print Dumper($info);
(specifically, the COLUMN_NAME and TYPE_NAME attributes).
However, this introduces a race condition if the table changes schema between the two queries.
Also, you may use selectall_arrayref with the Slice parameter to fetch all the columns into a hash ref, it needs no prepared statement and will return an array ref of the result set rows, with each rows columns the key's to a hash and the values are the column values. ie:
my $result = $dbh->selectall_arrayref( qq{
SELECT * FROM table WHERE condition = value
}, { Slice => {} }) or die "Error: ".$dbh->errstr;
$result = [
[0] = { column1 => 'column1Value', column2 => 'column2Value', etc...},
[1] = { column1 => 'column1Value', column2 => 'column2Value', etc...},
];
Making it easy to iterate over results.. ie:
for my $row ( #$results ){
print "$row->{column1Value}, $row->{column2Value}\n";
}
You can also specify which columns to extract but it's pretty useless due to the fact it's more efficient to do that in your SQL query syntax.
{ Slice => { column1Name => 1, column2Name => 1 } }
That would only return the values for column1Name and column2Name just like saying in your SQL:
SELECT column1Name, column2Name FROM table...

Perl DBI insert and select

I want to copy a single record from a table, modify some of the fields in the record and insert into the same table. The table has 90 columns.
Thought of using insert..select in one statement but there are 90 columns and i need to tell the column name in the select query. How can i do it in a better way in perldbi. Pls provide me an example.
Fetch and cache the column names for your table using the NAME statement attribute
my $sth = $dbh->prepare('SELECT * FROM my_table where 1=0');
$sth->execute;
my $cols = $sth->{NAME};
then use $cols to construct your insert...select using some replacement function to inject your modifications in the select.
my %mods_for_column = ( 'foo' => 'foo+10', 'bar' => 'trim(bar)' );
my $inscols = join(',', #$cols);
my $selcols = join(',',
map { exists($mods_for_column($_)) ? $mods_for_column($_) : $_ } #$cols
);
my $sql = "insert into my_table ($inscols) select $selcols from my_table where mumble...";

How do I insert values from parallel arrays into a database using Perl's DBI module?

I need to insert values in database using Perl's DBI module. I have parsed a file to obtain these values and hence these values are present in an arrays, say #array1, #array2, #array3. I know how to insert one value at a time but not from an arrays.
I know insert one value at a time:
$dbh = DBI->connect("dbi:Sybase:server=$Srv;database=$Db", "$user", "$passwd") or die "could not connect to database";
$query= "INSERT INTO table1 (id, name, address) VALUES (DEFAULT, tom, Park_Road)";
$sth = $dbh->prepare($query) or die "could not prepare statement\n";
$sth-> execute or die "could not execute statement\n $command\n";
I am not sure if I have array1 containing ids, array2 containing names, and array3 containing address, how would I insert values.
Since you have parallel arrays, you could take advantange of execute_array:
my $sth = $dbh->prepare('INSERT INTO table1 (id, name, address) VALUES (?, ?, ?)');
my $num_tuples_executed = $sth->execute_array(
{ ArrayTupleStatus => \my #tuple_status },
\#ids,
\#names,
\#addresses,
);
Please note that this is a truncated (and slightly modified) example from the documentation. You'll definitely want to check out the rest of it if you decide to use this function.
Use placeholders.
Update: I just realized you have parallel arrays. That is really not a good way of working with data items that go together. With that caveat, you can use List::MoreUtils::each_array:
#!/usr/bin/perl
use strict; use warnings;
use DBI;
use List::MoreUtils qw( each_array );
my $dbh = DBI->connect(
"dbi:Sybase:server=$Srv;database=$Db",
$user, $passwd,
) or die sprintf 'Could not connect to database: %s', DBI->errstr;
my $sth = $dbh->prepare(
'INSERT INTO table1 (id, name, address) VALUES (?, ?, ?)'
) or die sprintf 'Could not prepare statement: %s', $dbh->errstr;
my #ids = qw( a b c);
my #names = qw( d e f );
my #addresses = qw( g h i);
my $it = each_array(#ids, #names, #address);
while ( my #data = $it->() ) {
$sth->execute( #data )
or die sprintf 'Could not execute statement: %s', $sth->errstr;
}
$dbh->commit
or die sprintf 'Could not commit updates: %s', $dbh->errstr;
$dbh->disconnect;
Note that the code is not tested.
You might also want to read the FAQ: What's wrong with always quoting "$vars"?.
Further, given that the only way you are handling error is by dying, you might want to consider specifying { RaiseError => 1 } in the connect call.
How could you not be sure what your arrays contain? Anyway the approach would be the iterate through one array and assuming the other arrays have corresponding values put those into the insert statement
Another way would be to use a hash as an intermediate storage area. IE:
my $hash = {};
foreach(#array1) {
$hash->{id} = $array1[$_];
$hash->{name} = $array2[$_];
$hash->{address} = $array3[$_];
}
foreach( keys %$hash ) {
$sql = "insert into table values(?,?,?)";
$sth = $dbh->prepare($sql) or die;
$sth->execute($hash->{id}, $hash->{name}, $hash->{address}) or die;
}
Though again this depends on the three arrays being synced up. However you could modify this to do value modifications or checks or greps in the other arrays within the first loop through array1 (ie: if your values in array2 and array3 are maybe stored as "NN-name" and "NN-address" where NN is the id from the first array and you need to find the corresponding values and remove the NN- with a s// regex). Depends on how your data is structured though.
Another note is to check out Class::DBI and see if it might provide a nicer and more object oriented way of getting your data in.