Database Reference
In-Depth Information
Comparing Conventional Path and Direct Path Loads
As mentioned previously, the direct path load can be faster than the conventional path load.
The conventional path load should be used in the following situations:
Access to the target table is required concurrently with the data load.
Inserts or updates will occur to a non-indexed target table concurrently with the
data load.
When loading data into a clustered table.
When loading a relatively small number of rows into a large table that has referential
and column-check integrity constraints or indexes.
Parallel Direct Path
A parallel direct path load allows multiple direct path load sessions to concurrently load
into the same data segments, by using the PARALLEL parameter on the command line or
in a parameter file.
sqlldr USERID=hr CONTROL=hrloader1.ctl DIRECT=TRUE PARALLEL=TRUE
sqlldr USERID=hr CONTROL=hrloader2.ctl DIRECT=TRUE PARALLEL=TRUE
sqlldr USERID=hr CONTROL=hrloader3.ctl DIRECT=TRUE PARALLEL=TRUE
SQL*Loader Express Mode
As mentioned earlier, SQL*Loader express mode is a quick way to load by specifying only
a table name on the command line. It can be used when the table columns are all number,
datetime, or character datatypes and the input data files contain only delimited character
data, for example:
C:\Temp>type countries.dat
UR,Uruguay,2
TK,Turkey,4
SY,Syria,4
NP,Nepal,3
MY,Myanmar,3
BG,Bangladesh,3
VT,The Vatican,1
A control file is not used. SQL*Loader uses the target table column definitions to match
up the datatypes and order of the input fields. You can override default values by using
parameters and values on the command line.
To run SQL*Loader in express mode, type in sqlldr followed by the username parameter
and TABLE=<target table> at the command prompt.
C:\Temp>
C:\Temp>sqlldr hr/hr@LNE6 table=countries
Search WWH ::




Custom Search