Hi,

While testing “file_fdw: Support multi-line HEADER option”, I noticed a small 
issue.

The doc says that the HEADER option cannot be used with the “binary" or “json" 
format:
```
   <varlistentry id="sql-copy-params-header">
    <term><literal>HEADER</literal></term>
    <listitem>
     <para>
      On output, if this option is set to <literal>true</literal>
      (or an equivalent Boolean value), the first line of the output will
      contain the column names from the table.
      Integer values <literal>0</literal> and <literal>1</literal> are
      accepted as Boolean values, but other integers are not allowed for
      <command>COPY TO</command> commands.
     </para>
     <para>
      On input, if this option is set to <literal>true</literal>
      (or an equivalent Boolean value), the first line of the input is
      discarded.  If set to a non-negative integer, that number of
      lines are discarded.  If set to <literal>MATCH</literal>, the first line
      is discarded, and it must contain column names that exactly match the
      table's columns, in both number and order; otherwise, an error is raised.
      The <literal>MATCH</literal> value is only valid for
      <command>COPY FROM</command> commands.
     </para>
     <para>
      This option is not allowed when using <literal>binary</literal> or 
<literal>json</literal> format.
     </para>
    </listitem>
   </varlistentry>
```

However, when I specified "header ‘0", the command did not fail. That means the 
current behavior depends on the value of the “header" option, not on presence:
```
evantest=# create foreign table ft (i int) server fs options (format 'binary', 
filename '/tmp/ft.bin',  header '1');
ERROR:  cannot specify HEADER in BINARY mode
evantest=# create foreign table ft (i int) server fs options (format 'binary', 
filename '/tmp/ft.bin',  header '0');
CREATE FOREIGN TABLE
```

As we can see, "header 1" fails, but header 0" is silently accepted. I don't 
think this behavior matches what the documentation describes.

For comparison, VACUUM has a similar option. “BUFFER_USAGE_LIMIT" is not 
allowed with "VACUUM FULL", and a value of 0 means disabling the buffer access 
strategy:
```
   <varlistentry>
    <term><literal>BUFFER_USAGE_LIMIT</literal></term>
    <listitem>
     <para>
      Specifies the
      <glossterm linkend="glossary-buffer-access-strategy">Buffer Access 
Strategy</glossterm>
      ring buffer size for <command>VACUUM</command>.  This size is used to
      calculate the number of shared buffers which will be reused as part of
      this strategy.  <literal>0</literal> disables use of a
      <literal>Buffer Access Strategy</literal>.  If <option>ANALYZE</option>
      is also specified, the <option>BUFFER_USAGE_LIMIT</option> value is used
      for both the vacuum and analyze stages.  This option can't be used with
      the <option>FULL</option> option except if <option>ANALYZE</option> is
      also specified.  When this option is not specified,
      <command>VACUUM</command> uses the value from
      <xref linkend="guc-vacuum-buffer-usage-limit"/>.  Higher settings can
      allow <command>VACUUM</command> to run more quickly, but having too
      large a setting may cause too many other useful pages to be evicted from
      shared buffers.  The minimum value is <literal>128 kB</literal> and the
      maximum value is <literal>16 GB</literal>.
     </para>
    </listitem>
   </varlistentry>
```

Using BUFFER_USAGE_LIMIT 0 with FULL is still rejected:
```
evantest=# vacuum (full, BUFFER_USAGE_LIMIT 0) t;
ERROR:  BUFFER_USAGE_LIMIT cannot be specified for VACUUM FULL
```

So VACUUM rejects BUFFER_USAGE_LIMIT based on the presence of the option, not 
its value. I think we should keep the behavior consistent here, and VACUUM's 
behavior better matches the documentation. Otherwise, I am afraid this could 
encourage more inconsistencies in the future.

The fix is straightforward. Since we already have the “header_specified" 
variable to indicate whether the option is present, we can check 
“header_specified" instead.

I reported a similar issue for the COPY command earlier in thread [1]. If this 
patch is accepted, then that one may be worth considering as well.

[1] 
https://www.postgresql.org/message-id/C1D2509E-E5D1-46B0-932C-B57AA7B963A1%40gmail.com

--
Chao Li (Evan)
HighGo Software Co., Ltd.
https://www.highgo.com/




Attachment: v1-0001-Reject-HEADER-with-binary-and-json-COPY-formats-b.patch
Description: Binary data

Reply via email to