INFILE
This option is valid only when DIRECT YES is in effect.
The following table lists the alternative sources of input data that you can use, and tells you which INFILE option to use for each input type:
Alternative sources of BMC AMI Unload input
Input type | INFILE option |
---|---|
Full (including Inline Image copies) or incremental image copy data sets | |
Cabinet copies created by BMC AMI Copy, the BMC Recovery Management for Db2 solution, or the BMC AMI Recovery for Db2 solution | |
DSN1COPY sequential data sets | |
(BMC.DB2.SPE2410) VSAM FlashCopy image copies | |
(BMC.DB2.SPE2410) VSAM linear data set | |
(BMC.DB2.SPE2410) VSAM FlashCopy image copies |
Restrictions
The following additional restrictions apply to the INFILE option:
- You must specify +SVRMODETAPE(I) when specifying a JCL-allocated, INFILE ddname using tape and unloading to an OUTPUT statement using tape.
- You cannot specify FILTERPART YES when unloading from image copies.
- BMC AMI Unload does not unload from LOB or XML copy data sets. However, you can unload from a copy of the base table if you do not select any LOB or XML columns.
- You cannot use a DDLIN data set for copy file input.
- You cannot use a concatenation of image copies for different tables when specifying INFILE ddname.
- You cannot use the following types of copy input:
- Encrypted copies created by BMC AMI Copy
- Instant Snapshot copies created by BMC AMI Copy
- Online consistent copies created by the BMC Recovery Management for Db2 or BMC Recovery for Db2 solutions
- BMC AMI Unload does not support unloading from PBR2 inline copies which contain more than one partition.
Additional considerations
The following additional considerations apply to the INFILE option:
- Unpredictable results might occur if the data sets do not match the options that you specify:
- If the input data set is an inline image copy, specify the INLINE YES option. If INLINE YES or the default, INLINE NO, does not match the type of input data set, rows might be omitted from the output.
- If the input data set is an incremental image copy, specify the INCREMENTAL option.
- If you are unloading all partitions of a partitioned table space, ensure that the data set includes a full image copy of all of the partitions.
- If you are unloading a specific partition, ensure that the input provides the specified partition with the full image copy data set either by including a DSNUM ALL copy, or a partlevel copy of the partition(s) specified in the UNLOAD command.
- When you specify INFILE to unload any of the following types of copy data sets, all selected rows must be at the current version (as defined in SYSIBM.SYSTABLES):
- An incremental image copy
- Any image copy created with the SYSTEMPAGES NO copy option.
IMAGECOPY
Specify INFILE IMAGECOPY to have BMC AMI Unload use the specified full or incremental image copy data set that is in the SYSIBM.SYSCOPY table or the BMCXCOPY table; BMCXCOPY applies only if you are unloading a cabinet copy.
For information that applies to all INFILE options, see Restrictions and Additional considerations.
Partitioned table spaces
If the table space is partitioned, BMC AMI Unload uses one of the following copy types:
- (BMC.DB2.SPE2304) Partition range image copies (duplicate SYSCOPY entries where DSNUM is not equal to 0). Each partition of the table must exist in one of the image copy files.
- (BMC.DB2.SPE2304) Individual partition image copies (unique SYSCOPY entries where DSNUM is not equal to 0) There must be an individual copy for each partition in the table.
- If there is not one copy for each partition that you unload, the first image copy that fulfills your specifications and that contains all partitions (DSNUM is equal to 0) will be used.
- (BMC.DB2.SPE2304) Inline copies produced using Reorg ICLIMIT are also considered.
For a partitioned table space:
- BMC AMI Unload uses the latest full or incremental copy to complete unload processing.
- If the latest copy is a DSNUM 0 copy, BMC AMI Unload uses a copy of the entire partitioned table space.
- If the latest copy is not a DSNUM 0 copy, BMC AMI Unload uses the partition-level copies, provided that:
- Every partition has a copy.
- For INLINE IMAGE COPY files, each copy has the same START_RBA value.
If these conditions are not met, the unload job fails.
Non-partitioned table spaces
BMC AMI Unload can unload multi-data-set, non-partitioned table spaces only from a single image copy that contains all of the individual data sets. You cannot unload non-partitioned table spaces from separate image copy data sets where DSNUM is not equal to 0.
Operands
FULL | (default) BMC AMI Unload unloads from a full image copy. BMC AMI Unload unloads from the most recent full image copy found in the SYSIBM.SYSCOPY table or the BMCXCOPY table; BMCXCOPY applies only if you are unloading a cabinet copy. To unload from a full image copy before the most recent one, specify a negative integer to count the specified number of full image copies back from the most recent one. |
---|---|
INCREMENTAL | BMC AMI Unload unloads from an incremental image copy. When unloading from an incremental image copy, BMC AMI Unload performs the following actions:
BMC AMI Unload unloads the most recent incremental image copy found in the SYSIBM.SYSCOPY table or the BMCXCOPY table; BMCXCOPY applies only if you are unloading a cabinet copy. To unload from an incremental image copy before the most recent one, specify a negative integer to count the specified number of incremental image copies back from the most recent one. BMC AMI Unload cannot use the INFILE option to unload an incremental image copy that does not contain dictionary pages and does contain compressed data. |
0 | (default) This option tells BMC AMI Unload that the specified file is the most recent image copy |
–integer | This option tells BMC AMI Unload to count the specified number of image copies back from the most recent one |
ddname
Use this option to specify a ddname or ddname prefix when unloading from a DSN1COPY sequential data set or a specific full or incremental image copy.
If you specify INFILE ddname, BMC AMI Unload unloads all rows from the specified data set that match your SELECT statement criteria.
Additional restrictions and considerations
In addition to the general restrictions and considerations on the INFILE option, the following restrictions and considerations apply to INFILE ddname:
- The following Db2 catalog information must match the copy that you are unloading:
- The database identifier (DBID) and page set identifier (PSID) of the database and table space that contain the specified tables
- The OBIDs of the tables that you specify, unless you specify the OBID for the table by using the OBID option
- The table definitionsHowever, if you added more columns to a table since the copy was made and you are unloading those columns, the new columns contain their default values.
- Concatenating multiple image copy data sets, particularly when mixing copy data sets from tape with copy data sets from DASD, produces unpredictable results.
- Using multiple SELECT statements for different tables with concatenated image copy data sets (using INFILE ddname) can cause failures or might unload the wrong data.
- When unloading an image copy on a Db2 subsystem other than the one on which the copy was created, the following information applies:
- For a range-partitioned table space, BMC AMI Unload unloads only the number of partitions that are defined on the target subsystem.
- For a partition-by-growth table space, BMC AMI Unload unloads only the number of partitions that exist on the target subsystem.
Operands
FULL | (Default) Tells BMC AMI Unload that the specified file is a full image copy or a DSN1COPY data set | ||||
---|---|---|---|---|---|
INLINE | BMC AMI Unload INFILE ddname FULL now supports unloading from an INLINE image copy using INFILE ddname FULL INLINE YES.
| ||||
INCREMENTAL | Tells BMC AMI Unload that the specified file is an incremental image copy If INFILE ddname specifies an incremental image copy, you must use this option or BMC AMI Unload assumes that FULL (the default) applies and unpredictable results might occur. When using an incremental image copy, BMC AMI Unload reads all rows in that image copy, not only the rows that changed. INFILE ddname INCREMENTAL does not support INLINE YES. | ||||
INFILEPFX | Tells BMC AMI Unload to scan for numbered ddnames (for example, ddn01, ddn02, ddn03, and so on) for a partitioned-by-range (PBR) table spaceBMC AMI Unload then serially reads these numbered ddnames instead of reading a single INFILE ddname. The numeric suffix can contain an arbitrary number of leading zeros for example, ddn0006, ddn7, or ddn008. The number of these ddnames corresponds to the number of selected partitions in the PBR table space. |
VSAMDD
Specify this option to unload from a VSAM linear data set or VSAM FlashCopy image copy data set. You must specify the data set as a VSAMDD DD statement in your JCL.
When unloading multiple partitions the INFILEPFX should be used along with the VSAMDD keyword. Using INFILEPFX tells BMC AMI Unload to scan for numbered VSAMDD data sets. The numeric suffix can contain an arbitrary number of leading zeros, for example, VSAMDD02 or VSAMDD2. The number of these ddnames corresponds to the number of unloading partitions in the table space.
With this option, you can specify up to 99 VSAM data sets for partitioned objects. If you need to unload from more than 99 data sets, use the VSAMDDPREFIX option.
VSAMDDPREFIX prefix
Specify this option to use a prefix other than VSAMDD when unloading from a VSAM linear data set or a VSAM FlashCopy image copy data set. This option enables users to unload from more than 99 data sets. When unloading multiple partitions the INFILEPFX should be used along with the VSAMDDPREFIX keyword.
The JCL must include DD statements that match the prefix that is specified. For multiple data sets, the user must append n to the ddnames in the DD statements in the JCL, where n is a numeric value. These ddnames must not exceed eight characters.