Edit report at https://bugs.php.net/bug.php?id=80874&edit=1
ID: 80874
Updated by: cmb@php.net
Reported by: tony at tonymarston dot net
Summary: odbc_fetch_array() will not fetch a row containing a
UTF-8 currency code symbol
-Status: Assigned
+Status: Feedback
Type: Bug
Package: ODBC related
Operating System: Windows 10
PHP Version: 7.4.16
Assigned To: cmb
Block user comment: N
Private report: N
New Comment:
> [â¦] and 'EFBBBF3F' in hexadecimal. Why does it show up as '?'
Because that bytes are an UTF-8 BOM followed by a question mark.
Displaying this as ? is correct, but I wonder where that BOM comes
from.
> I have CHAR_AS_UTF8=TRUE in the additional connection properties
> in the ODBC Data Source Administrator.
Did you try with CHAR_AS_UTF8=FALSE?
Also, I suggest that you check whether that would work with
PDO_ODBC. I assume that you'd get the same results as with ODBC by
default, but setting the PDO::ODBC_ATTR_ASSUME_UTF8 attribute may
yield the desired results.
Previous Comments:
------------------------------------------------------------------------
[2021-03-18 15:28:27] tony at tonymarston dot net
No notice or warning is generated on the call to odbc_fetch_array().
I have CHAR_AS_UTF8=TRUE in the additional connection properties in the ODBC Data Source
Administrator.
When I access my SQL Server database via ODBC the currency symbol returned for THB appears as
'?' in the PHP array and also when I display it in the HTML output. However, when I copy
it to my text editor it is shown as '?' in ASCII and 'EFBBBF3F' in
hexadecimal. Why does it show up as '?'
------------------------------------------------------------------------
[2021-03-18 13:03:49] cmb@php.net
>> Doesn't it also raise a warning?
> How do I detect that? There is no ODBC function to retrieve
> warnings.
I was referring to the general PHP error reporting mechanisms.
Anyway, the ODBC trace shows whats going on:
| General error;-10427 Conversion of parameter/column (3) from
| data type NVARCHAR to ASCII failed (-10427)
Conversion to ASCII can't succeed, but PHP does not enforce that,
but rather requests binding to SQL_C_CHAR. Is there a respective
setting to convert to UTF-8 in the driver options?
With SQL Server (ODBC Driver 17 for SQL Server) all currency
symbols are properly retrieved and displayed for me. Watch out
for font issues (call bin2hex() to see the byte values).
------------------------------------------------------------------------
[2021-03-17 18:27:28] tony at tonymarston dot net
> Doesn't it also raise a warning?
How do I detect that? There is no ODBC function to retrieve warnings.
The trace log can be downloaded from http://www.tonymarston.net/odbc.log
I have also tried running the same test on my SQL Server database, but while it does return all 9
rows the currency_symbol column for 'THB' and 'YUAN' are being returned as
'?' instead of the expected UTF-8 symbol.
------------------------------------------------------------------------
[2021-03-17 16:02:18] cmb@php.net
> [â¦] but the call to odbc_fetch_array() only returns the first 2
> rows and returns FALSE for the 3rd row.
Doesn't it also raise a warning?
Anyhow, I cannot reproduce this (I haven't a SAP HANA DB
available, though), so please provide an ODBC trace so we can see
what's actually happening. If possible, put it somewhere on the
Web (e.g. gist.github.com, pastebin.com) and only post a link to
the trace here.
------------------------------------------------------------------------
[2021-03-16 18:11:38] tony at tonymarston dot net
Description:
------------
I have a database table in a SAP HANA database which contains currency codes and their symbols, but
when I try to read them in my PHP program the rows containing symbols are not returned in the result
set.
Test script:
---------------
CREATE TABLE currency (
currency_code nvarchar(16) NOT NULL,
currency_name nvarchar(255) NOT NULL,
currency_symbol nvarchar(4) DEFAULT NULL,
PRIMARY KEY (currency_code)
);
INSERT INTO currency (currency_code, currency_name, currency_symbol) VALUES ('AUD',
'Australian Dollars', NULL);
INSERT INTO currency (currency_code, currency_name, currency_symbol) VALUES ('CAD',
'Canadian Dollars', NULL);
INSERT INTO currency (currency_code, currency_name, currency_symbol) VALUES ('EUR',
'Euros', 'â¬');
INSERT INTO currency (currency_code, currency_name, currency_symbol) VALUES ('GBP',
'UK Pounds', '£');
INSERT INTO currency (currency_code, currency_name, currency_symbol) VALUES ('JPY',
'Japanese Yen', 'Â¥');
INSERT INTO currency (currency_code, currency_name, currency_symbol) VALUES ('SGD',
'Singapore Dollar', '$');
INSERT INTO currency (currency_code, currency_name, currency_symbol) VALUES ('THB',
'Thai Baht', '฿');
INSERT INTO currency (currency_code, currency_name, currency_symbol) VALUES ('USD',
'US Dollars', '$');
INSERT INTO currency (currency_code, currency_name, currency_symbol) VALUES ('YUAN',
'Chinese Yuan Renminbi', 'å
');
<?php
$driver = 'HDBODBC';
$server = 'hxehost:39015';
$user = '****';
$pswd = '****';
$conn = odbc_connect("Driver=$driver;ServerNode=$server",$user,$pswd) OR
die('Connection to DB via ODBC failed');
$query = "SELECT * FROM test.currency ORDER BY currency_code";
if (!$result = odbc_exec($conn, $query)) {
die("ERRORNO=".odbc_error().", ERRORMSG=".odbc_errormsg());
};
$numrows = odbc_num_rows($result);
while ($row = odbc_fetch_array($result)) {
$array[] = array_change_key_case($row, CASE_LOWER);
echo "<p>id={$row['CURRENCY_CODE']},
name={$row['CURRENCY_NAME']},
symbol={$row['CURRENCY_SYMBOL']}</p>\n";
} // while
odbc_free_result($result);
?>
Expected result:
----------------
The call to odbc_num_rows() returns '9' which is correct, but the call to
odbc_fetch_array() only returns the first 2 rows and returns FALSE for the 3rd row.
------------------------------------------------------------------------
--
Edit this bug report at https://bugs.php.net/bug.php?id=80874&edit=1