版权声明：如若转载或他用，望告知 https://blog.csdn.net/qwerdfcv/article/details/80619373
导出数据库表结构对象第一步，便是要导出对应的数据表内容，首先给出一段导出表字段描述的SQL：
DECLARE
CURSOR cols(p_table_name VARCHAR2, p_owner VARCHAR2) IS
SELECT tc.column_id,
tc.column_name,
cc.comments,
tc.data_type,
tc.data_length,
decode(tc.nullable, 'N', '', '√') AS nullable,
tc.data_default
FROM dba_tab_columns /*sys.all_tab_columns*/ tc, dba_col_comments /*sys.all_col_comments*/ cc
WHERE tc.table_name = upper(p_table_name)
AND tc.owner = upper(p_owner)
AND tc.owner = cc.owner
AND tc.table_name = cc.table_name
AND tc.column_name = cc.column_name
ORDER BY column_id;
lv_table_name VARCHAR2(30) := 'table_name';
lv_owner VARCHAR2(10) := 'user';
BEGIN
dbms_output.put_line('<html><table><thead><th>序号</th><th>字段名称</th><th>字段描述</th><th>字段类型</th><th>长度</th><th>允许空</th><th>缺省值</th></thead>');
FOR c IN cols(lv_table_name, lv_owner) LOOP
dbms_output.put_line('<tr><td>' || c.column_id || '</td><td>' || c.column_name || '</td><td>' ||
c.comments || '</td><td>' || c.data_type || '</td><td>' || c.data_length ||
'</td><td>' || c.nullable || '</td><td>' || c.data_default ||
'</td></tr>');
END LOOP;
dbms_output.put_line('</table></br></html>');
END;
接下来再对此SQL进行一下优化，使其导出的内容更加具体一点，包含表设计具体内容：
DECLARE
CURSOR cols(p_table_name VARCHAR2, p_owner VARCHAR2) IS
SELECT tc.column_id,
tc.column_name,
cc.comments,
tc.data_type,
tc.data_length,
decode(tc.nullable, 'N', '', '√') AS nullable,
tc.data_default
FROM dba_tab_columns /*sys.all_tab_columns*/ tc, dba_col_comments /*sys.all_col_comments*/ cc
WHERE tc.table_name = upper(p_table_name)
AND tc.owner = upper(p_owner)
AND tc.owner = cc.owner
AND tc.table_name = cc.table_name
AND tc.column_name = cc.column_name
ORDER BY column_id;
lv_table_name VARCHAR2(30) := 'table_name';
lv_owner VARCHAR2(10) := 'user';
BEGIN
dbms_output.put_line('<html><table><thead><th>序号</th><th>字段名称</th><th>字段描述</th><th>字段类型</th><th>长度</th><th>允许空</th><th>缺省值</th></thead>');
FOR c IN cols(lv_table_name, lv_owner) LOOP
dbms_output.put_line('<tr><td>' || c.column_id || '</td><td>' || c.column_name || '</td><td>' ||
c.comments || '</td><td>' || c.data_type || '</td><td>' || c.data_length ||
'</td><td>' || c.nullable || '</td><td>' || c.data_default ||
'</td></tr>');
END LOOP;
dbms_output.put_line('</table></br></html>');
END;
好了，这样一张表的导出，基本上已经符合我们的需求了，那么，接下来就是需要给他加点样式，让后导出整个数据库用户对应的所有表结构了：
/*导出 TABLE 数据库对象的描述*/
/*下面是html版本,把DBMS输出存为doc或htm或xls文件即可.*/
/*注意对不同Schema需要修改此owner*/
DECLARE
CURSOR tabs IS
SELECT tab.table_name,
com.comments,
tab.owner,
tab.tablespace_name AS space_name,
tab.status AS status,
decode(tab.logging, 'NO', '', '√') AS logged,
tab.num_rows AS num_rows,
round(((tab.blocks * 8192) / 1024 / 1024 / 1024), 8) AS val,
to_char(tab.last_analyzed, 'yyyy/mm/dd hh24:mi:ss') AS last_analyzed,
decode(tab.partitioned, 'NO', '', '√') AS partitioned
FROM dba_tables /*sys.ALL_ALL_TABLES*/ tab, dba_tab_comments /*sys.all_tab_comments*/ com
WHERE tab.table_name = com.table_name
AND tab.owner = com.owner
AND tab.table_name NOT LIKE 'TOAD_%'
AND tab.table_name NOT LIKE 'TEMP_%'
AND tab.table_name NOT LIKE 'TEST_%'
AND tab.table_name NOT LIKE 'TMP_%'
AND tab.table_name NOT LIKE '%MLOG_%'
AND tab.table_name NOT LIKE '%SQLN_%'
AND tab.table_name NOT LIKE '%$%'
AND tab.table_name NOT LIKE '%PLSQL_%'
AND tab.table_name NOT LIKE '%_BK%'
AND tab.table_name NOT LIKE '%BIG_DEL%'
AND tab.table_name NOT LIKE '%_091021%'
AND tab.table_name NOT LIKE '%BAK%'
AND tab.table_name NOT LIKE '%MY_%'
AND tab.table_name NOT LIKE '%WKZ%'
AND tab.table_name NOT LIKE 'BIG%'
AND tab.table_name NOT IN ('TBLMITEMCCLASS2MITEM_BF080920',
'TBLPARTDEVICE_091024',
'TBLUGU20111229BACK',
'TBLIQCCHECKCRITERIONTMP1028',
'QUEST_SL_TEMP_EXPLAIN1',
'TEST',
'MIS2012TMP',
'WKZTEMP0504')
AND (tab.owner = upper('user'))
ORDER BY tab.table_name;
CURSOR cols(p_table_name VARCHAR2, p_owner VARCHAR2) IS
SELECT tc.column_id,
tc.column_name,
cc.comments,
tc.data_type,
tc.data_length,
decode(tc.nullable, 'N', '', '√') AS nullable,
tc.data_default
FROM dba_tab_columns /*sys.all_tab_columns*/ tc, dba_col_comments /*sys.all_col_comments*/ cc
WHERE tc.table_name = p_table_name
AND tc.owner = p_owner
AND tc.owner = cc.owner
AND tc.table_name = cc.table_name
AND tc.column_name = cc.column_name
ORDER BY column_id;
v_css VARCHAR2(4000);
v_tab_header VARCHAR2(2000);
v_tab_footer VARCHAR2(20);
v_number NUMBER DEFAULT 1;
BEGIN
v_css := '<html>
<style type="text/css">
h2
{
font-size: 15pt;
color: #4F81BD;
font-family: "微软雅黑";
}
h5
{
font-size: 13pt;
color: #4F81BD;
font-family: "微软雅黑";
}
.tn
{
font-size: 13pt;
color: white;
background-color:#4F81BD;
font-family: "微软雅黑";
}
table
{
font-size: 12pt;
font-family: "微软雅黑";
width: 100%;
border-collapse: collapse;
border: none;
padding-left: 5pt;
mso-border-alt: solid #4F81BD 1.0pt;
mso-padding-alt: 0cm 5.4pt 0cm 5.4pt;
}
thead
{
mso-yfti-irow: 0;
font-family: "微软雅黑";
mso-yfti-firstrow: yes;
}
th
{
font-family: "微软雅黑";
font-size: 12pt;
padding: 6pt 6pt 6pt 6pt;
border-top: solid #4F81BD 1.0pt;
border-left: solid #4F81BD 1.0pt;
border-bottom: solid #4F81BD 3.0pt;
border-right: solid #4F81BD 1.0pt;
}
tr
{
mso-yfti-irow: 1;
}
td
{
text-align: justify;
text-justify: inter-ideograph;
font-size: 10pt;
font-family: "微软雅黑";
margin: 0pt;
padding: 4pt 4pt 4pt 4pt;
border-top: none;
border-left: solid #4F81BD 1.0pt;
border-bottom: solid #4F81BD 1.0pt;
border-right: solid #4F81BD 1.0pt;
}
</style>';
dbms_output.put_line(v_css);
dbms_output.put_line('<table><thead><th>序号</th><th>表名</th><th>描述</th><th>表空间</th><th>状态</th><th>日志</th><th>表行数</th><th>表容量(G)</th><th>上次分析时间</th><th>是否分区</th></thead>');
FOR t IN tabs LOOP
dbms_output.put_line('<tr><td align=center>' || lpad(v_number, 3, 0) ||
'</td><td><a class="tn" href="#' || t.table_name || '">' || t.table_name ||
'</a></td><td>' || t.comments || '</td><td>' || t.space_name ||
'</td><td>' || t.status || '</td><td>' || t.logged || '</td><td>' ||
t.num_rows || '</td><td>' || t.val || '</td><td>' || t.last_analyzed ||
'</td><td>' || t.partitioned || '</td></tr>');
v_number := v_number + 1;
END LOOP;
v_tab_footer := '</table></br>';
dbms_output.put_line(v_tab_footer);
v_tab_header := '<table><thead><th>序号</th><th>字段名称</th><th>字段描述</th><th>字段类型</th><th>长度</th><th>允许空</th><th>缺省值</th></thead>';
FOR t IN tabs LOOP
dbms_output.put_line('<h2><a name="' || t.table_name || '">' || t.table_name || '</a></h2>');
dbms_output.put_line('<h5>表描述:<font color="black">' || t.comments || '</font></h5>');
dbms_output.put_line(v_tab_header);
FOR c IN cols(t.table_name, t.owner) LOOP
dbms_output.put_line('<tr><td>' || c.column_id || '</td><td>' || c.column_name ||
'</td><td>' || c.comments || '</td><td>' || c.data_type || '</td><td>' ||
c.data_length || '</td><td>' || c.nullable || '</td><td>' ||
c.data_default || '</td></tr>');
END LOOP;
dbms_output.put_line(v_tab_footer);
END LOOP;
dbms_output.put_line('</html>');
END;
这样便能输出带样式的数据库表结构对象了。
注意：
一般来说，一个用户的拥有表有时会很多的，而Oracle数据库的输出限制是1000000 bytes，所以常会有超出限制的情况发生，当超出限制时，只需要再输出内容前执行此语句即可：
dbms_output.enable(buffer_size => NULL);