1.先看一例子
expdp aiki/\"KFC~!@\" ESTIMATE=statistics parfile=kfc.par --linux环境
$more kfc.par
directory=monica_dir
dumpfile=kfc.dmp
logfile=kfclog.log
exclude=table:"like 'PJ_%'"
涉及参数
estimate,查阅文档
Default: BLOCKS
Purpose
Specifies the method that Export will use to estimate how much disk space each table
in the export job will consume (in bytes). The estimate is printed in the log file and
displayed on the client's standard output device. The estimate is for table row data
only; it does not include metadata.
译:指定这个参数的目的是在导出的时候预估下将耗用的磁盘空间（字节），预估的信息将显示在客户的显示器或输出设备上。
预估只针对表行的数据，不针对元数据；--元数据又是什么？google->描述数据的数据，如段头描述这个表有多少行，物理地址啥的。。
Syntax and Description
ESTIMATE={BLOCKS | STATISTICS}
■ BLOCKS - The estimate is calculated by multiplying the number of database blocks
used by the source objects, times the appropriate block sizes.
第一个参数"BLOCKS"预估的值将按原来几倍于数据块大小*数据块数的值；
■ STATISTICS - The estimate is calculated using statistics for each table. For this
method to be as accurate as possible, all tables should have been analyzed
recently.
第二个参数"STATISTICS"将使用表的统计数据，所以为了保证显示数据的准确性，所有的导出表应该在最近分析过；
Restrictions
If the Data Pump export job involves compressed tables, the default size estimation
given for the compressed table is inaccurate when ESTIMATE=BLOCKS is used. This is
because the size estimate does not reflect that the data was stored in a compressed
form. To get a more accurate size estimate for compressed tables, use
ESTIMATE=STATISTICS.
限制:如果导出的是压缩表，当参数为blocks时它将不精确的，因为压缩表的不能预估时不能反应出其预估大小，
所以为了获得更精确压缩表的数据大小请使用statistics参数;
parfile
继续查阅文档
Default: none
Purpose
Specifies the name of an export parameter file.
Syntax and Description
PARFILE=[directory_path]file_name
Unlike dump and log files, which are created and written by the Oracle database, the
parameter file is opened and read by the client running the expdp image. Therefore, a
directory object name is neither required nor appropriate. The directory path is an
operating system-specific directory specification. The default is the user's current
directory.默认是不使用的，格式如"/home/oracle/kfc.par"[目录/file]
The use of parameter files is highly recommended if you are using parameters whose
values require the use of quotation marks. See Use of Quotation Marks On the Data
Pump Command Line on page 2-6.
Restrictions
The PARFILE parameter cannot be specified within a parameter file.
Example
The content of an example parameter file, hr.par, might be as follows:
SCHEMAS=HR
DUMPFILE=exp.dmp
DIRECTORY=dpump_dir1
LOGFILE=exp.log
You could then issue the following Export command to specify the parameter file:
> expdp hr/hr parfile=hr.par
