doc/src/sgml/dml.sgml

   1 <!-- $PostgreSQL$ -->
   2
   3 <chapter id="dml">
   4  <title>Data Manipulation</title>
   5
   6  <remark>
   7   This chapter is still quite incomplete.
   8  </remark>
   9
  10  <para>
  11   The previous chapter discussed how to create tables and other
  12   structures to hold your data.  Now it is time to fill the tables
  13   with data.  This chapter covers how to insert, update, and delete
  14   table data.  We also introduce ways to effect automatic data changes
  15   when certain events occur: triggers and rewrite rules.  The chapter
  16   after this will finally explain how to extract your long-lost data
  17   back out of the database.
  18  </para>
  19
  20  <sect1 id="dml-insert">
  21   <title>Inserting Data</title>
  22
  23   <indexterm zone="dml-insert">
  24    <primary>inserting</primary>
  25   </indexterm>
  26
  27   <indexterm zone="dml-insert">
  28    <primary>INSERT</primary>
  29   </indexterm>
  30
  31   <para>
  32    When a table is created, it contains no data.  The first thing to
  33    do before a database can be of much use is to insert data.  Data is
  34    conceptually inserted one row at a time.  Of course you can also
  35    insert more than one row, but there is no way to insert less than
  36    one row at a time.  Even if you know only some column values, a
  37    complete row must be created.
  38   </para>
  39
  40   <para>
  41    To create a new row, use the <xref linkend="sql-insert"
  42    endterm="sql-insert-title"> command.  The command requires the
  43    table name and a value for each of the columns of the table.  For
  44    example, consider the products table from <xref linkend="ddl">:
  45 <programlisting>
  46 CREATE TABLE products (
  47     product_no integer,
  48     name text,
  49     price numeric
  50 );
  51 </programlisting>
  52    An example command to insert a row would be:
  53 <programlisting>
  54 INSERT INTO products VALUES (1, 'Cheese', 9.99);
  55 </programlisting>
  56    The data values are listed in the order in which the columns appear
  57    in the table, separated by commas.  Usually, the data values will
  58    be literals (constants), but scalar expressions are also allowed.
  59   </para>
  60
  61   <para>
  62    The above syntax has the drawback that you need to know the order
  63    of the columns in the table.  To avoid that you can also list the
  64    columns explicitly.  For example, both of the following commands
  65    have the same effect as the one above:
  66 <programlisting>
  67 INSERT INTO products (product_no, name, price) VALUES (1, 'Cheese', 9.99);
  68 INSERT INTO products (name, price, product_no) VALUES ('Cheese', 9.99, 1);
  69 </programlisting>
  70    Many users consider it good practice to always list the column
  71    names.
  72   </para>
  73
  74   <para>
  75    If you don't have values for all the columns, you can omit some of
  76    them.  In that case, the columns will be filled with their default
  77    values.  For example:
  78 <programlisting>
  79 INSERT INTO products (product_no, name) VALUES (1, 'Cheese');
  80 INSERT INTO products VALUES (1, 'Cheese');
  81 </programlisting>
  82    The second form is a <productname>PostgreSQL</productname>
  83    extension.  It fills the columns from the left with as many values
  84    as are given, and the rest will be defaulted.
  85   </para>
  86
  87   <para>
  88    For clarity, you can also request default values explicitly, for
  89    individual columns or for the entire row:
  90 <programlisting>
  91 INSERT INTO products (product_no, name, price) VALUES (1, 'Cheese', DEFAULT);
  92 INSERT INTO products DEFAULT VALUES;
  93 </programlisting>
  94   </para>
  95
  96   <para>
  97    You can insert multiple rows in a single command:
  98 <programlisting>
  99 INSERT INTO products (product_no, name, price) VALUES
 100     (1, 'Cheese', 9.99),
 101     (2, 'Bread', 1.99),
 102     (3, 'Milk', 2.99);
 103 </programlisting>
 104   </para>
 105
 106   <tip>
 107    <para>
 108     When inserting a lot of data at the same time, considering using
 109     the <xref linkend="sql-copy" endterm="sql-copy-title"> command.
 110     It is not as flexible as the <xref linkend="sql-insert"
 111     endterm="sql-insert-title"> command, but is more efficient. Refer
 112     to <xref linkend="populate"> for more information on improving
 113     bulk loading performance.
 114    </para>
 115   </tip>
 116  </sect1>
 117
 118  <sect1 id="dml-update">
 119   <title>Updating Data</title>
 120
 121   <indexterm zone="dml-update">
 122    <primary>updating</primary>
 123   </indexterm>
 124
 125   <indexterm zone="dml-update">
 126    <primary>UPDATE</primary>
 127   </indexterm>
 128
 129   <para>
 130    The modification of data that is already in the database is
 131    referred to as updating.  You can update individual rows, all the
 132    rows in a table, or a subset of all rows.  Each column can be
 133    updated separately; the other columns are not affected.
 134   </para>
 135
 136   <para>
 137    To perform an update, you need three pieces of information:
 138    <orderedlist spacing="compact">
 139     <listitem>
 140      <para>The name of the table and column to update,</para>
 141     </listitem>
 142
 143     <listitem>
 144      <para>The new value of the column,</para>
 145     </listitem>
 146
 147     <listitem>
 148      <para>Which row(s) to update.</para>
 149     </listitem>
 150    </orderedlist>
 151   </para>
 152
 153   <para>
 154    Recall from <xref linkend="ddl"> that SQL does not, in general,
 155    provide a unique identifier for rows.  Therefore it is not
 156    necessarily possible to directly specify which row to update.
 157    Instead, you specify which conditions a row must meet in order to
 158    be updated.  Only if you have a primary key in the table (no matter
 159    whether you declared it or not) can you reliably address individual rows,
 160    by choosing a condition that matches the primary key.
 161    Graphical database access tools rely on this fact to allow you to
 162    update rows individually.
 163   </para>
 164
 165   <para>
 166    For example, this command updates all products that have a price of
 167    5 to have a price of 10:
 168 <programlisting>
 169 UPDATE products SET price = 10 WHERE price = 5;
 170 </programlisting>
 171     This might cause zero, one, or many rows to be updated.  It is not
 172     an error to attempt an update that does not match any rows.
 173   </para>
 174
 175   <para>
 176    Let's look at that command in detail. First is the key word
 177    <literal>UPDATE</literal> followed by the table name.  As usual,
 178    the table name can be schema-qualified, otherwise it is looked up
 179    in the path.  Next is the key word <literal>SET</literal> followed
 180    by the column name, an equals sign and the new column value.  The
 181    new column value can be any scalar expression, not just a constant.
 182    For example, if you want to raise the price of all products by 10%
 183    you could use:
 184 <programlisting>
 185 UPDATE products SET price = price * 1.10;
 186 </programlisting>
 187    As you see, the expression for the new value can refer to the existing
 188    value(s) in the row.  We also left out the <literal>WHERE</literal> clause.
 189    If it is omitted, it means that all rows in the table are updated.
 190    If it is present, only those rows that match the
 191    <literal>WHERE</literal> condition are updated.  Note that the equals
 192    sign in the <literal>SET</literal> clause is an assignment while
 193    the one in the <literal>WHERE</literal> clause is a comparison, but
 194    this does not create any ambiguity.  Of course, the
 195    <literal>WHERE</literal> condition does
 196    not have to be an equality test.  Many other operators are
 197    available (see <xref linkend="functions">).  But the expression
 198    needs to evaluate to a Boolean result.
 199   </para>
 200
 201   <para>
 202    You can update more than one column in an
 203    <command>UPDATE</command> command by listing more than one
 204    assignment in the <literal>SET</literal> clause.  For example:
 205 <programlisting>
 206 UPDATE mytable SET a = 5, b = 3, c = 1 WHERE a &gt; 0;
 207 </programlisting>
 208   </para>
 209  </sect1>
 210
 211  <sect1 id="dml-delete">
 212   <title>Deleting Data</title>
 213
 214   <indexterm zone="dml-delete">
 215    <primary>deleting</primary>
 216   </indexterm>
 217
 218   <indexterm zone="dml-delete">
 219    <primary>DELETE</primary>
 220   </indexterm>
 221
 222   <para>
 223    So far we have explained how to add data to tables and how to
 224    change data.  What remains is to discuss how to remove data that is
 225    no longer needed.  Just as adding data is only possible in whole
 226    rows, you can only remove entire rows from a table.  In the
 227    previous section we explained that SQL does not provide a way to
 228    directly address individual rows.  Therefore, removing rows can
 229    only be done by specifying conditions that the rows to be removed
 230    have to match.  If you have a primary key in the table then you can
 231    specify the exact row.  But you can also remove groups of rows
 232    matching a condition, or you can remove all rows in the table at
 233    once.
 234   </para>
 235
 236   <para>
 237    You use the <xref linkend="sql-delete" endterm="sql-delete-title">
 238    command to remove rows; the syntax is very similar to the
 239    <command>UPDATE</command> command.  For instance, to remove all
 240    rows from the products table that have a price of 10, use:
 241 <programlisting>
 242 DELETE FROM products WHERE price = 10;
 243 </programlisting>
 244   </para>
 245
 246   <para>
 247    If you simply write:
 248 <programlisting>
 249 DELETE FROM products;
 250 </programlisting>
 251    then all rows in the table will be deleted!  Caveat programmer.
 252   </para>
 253  </sect1>
 254 </chapter>