|
Column Name
|
Description
|
Specification
|
|---|---|---|
|
ModelName
|
contains the name of
the model.
|
NVARCHAR(128) NOT NULL
|
|
ModelSequence
|
contains the sequence
number for the block of ModelDS2 data.
The ModelDS2 files is
divided into blocks with each block in one row for one model.
|
INTEGER NOT NULL
|
|
ModelDS2
|
contains the sasscore_modelname.ds2
file.
|
NVARCHAR (8000)
|
|
ModelFormats
|
contains the sasscore_modelname_ufmt.xml
file.
|
NVARCHAR (8000)
|
|
ModelUUID1
|
contains the UUID of
the source model.
|
VARCHAR (36)
|
|
Notes1
|
contains additional
information that describes the source model.
|
VARCHAR (512)
|
| 1This column is for use by SAS Model Manager. If you have a model table that was created prior to SAS 9.4 and you want this column in your model table, you must run the %INDNZ_CREATE_MODELTABLEmacro to re-create your model table. | ||
%let indconn = server=myserver user=myuserid password=xxxx database=mydb;
specifies the name of a Netezza database where the sasscore_modelname.ds2 and sasscore_modelname_ufmt.xml scoring files are held.
| Default | The database specified in the INDCONN macro variable or your current database. |
specifies the name of the table that holds the sasscore_modelname.ds2 and sasscore_modelname_ufmt.xml scoring files.
| Default | sas_model_table |
| Requirement | The maximum table name length is 128 characters, and it must be a valid Netezza table name. |
| Interaction | The table name that you specify for this macro must be the same table name that is used in the %INDNZ_PUBLISH_MODEL macro. |
| See | %INDNZ_PUBLISH_MODEL Macro Syntax |
specifies one of the following actions that the macro performs:
creates a new table.
| Tip | If the table has been previously defined and you specify ACTION=CREATE, an error is issued. |
overwrites the current table, if a table with the same name is already registered.
| Tip | If you specify ACTION = REPLACE, and the current table contains sasscore_modelname.ds2 and sasscore_modelname_ufmt.xml files, the files are deleted and an empty table is re-created. |
causes all models in this table to be dropped.
| Default | CREATE |
proc sql; connect to netezza (user=userid password=xxxx server=myserver); execute (CALL NZRC..SAS_EP (' model_name=intr_reg, input_table=intr_reg_20m, input_columns= id count dif_srvr flag hot sam_srat service srv_cnt, output_table=intr_reg_out, journal_table=intr_reg_out_jnl, ds2_keep=id EM_CLASSIFICATION EM_EVENTPROBABILITY EM_PROBABILITY '); ) by netezza; disconnect from netezza; quit;
specifies the name of the model.
| Requirement | There must be at least one row in the model table with this key in the ModelName column and with a non-null value in the ModelDS2 column. |
| Tip | The model name is case sensitive. |
specifies the name of the scoring model input table.
| Requirement | If the input table name is case sensitive, you must use
double quotation marks (for example, "myinDB..myinTABLE").
|
specifies the name of the scoring model output table.
| Requirement | If the output table name is case sensitive, you must
use double quotation marks (for example, "myoutDB..myoutTABLE").
|
specifies the name of the model table where the sasscore_modelname.ds2 and sasscore_modelname_ufmt.xml scoring files were published with the %INDTD_CREATE_MODELTABLE macro.
| Default | sas_model_table |
| Restriction | This argument is available only when using the SAS Embedded Process. |
| Requirements | If the model table name is case sensitive, you must use
double quotation marks(for example, "mymtDB..mymtTABLE").
|
| The name of the model table must be the same as the name specified in the %INDNZ_CREATE_MODELTABLE macro. For more information, see the MODELTABLE argument in %INDNZ_CREATE_MODELTABLE Macro Syntax. |
specifies one or more columns from the input table that are passed to the SAS Embedded Process.
| Default | All columns |
| Requirement | If the column name is case sensitive, you must use double
quotation marks (for example, "Profit" "Sales"
"Margin").
|
specifies a valid WHERE clause that selects the rows from the input table.
specifies one or more columns that are the distribution key for the output table.
| Default | Current database distribution |
| Requirement | If the column name is case sensitive, you must use double
quotation marks (for example, "Profit", "Sales",
"Margin").
|
specifies whether the output table is temporary.
| Default | NO |
specifies the name of a table that the SAS_EP stored procedure creates. This table holds any journal messages and notes from the SAS journal facility that are produced when executing the store procedure.
| Requirement | If the journal table name is case sensitive, you must
use double quotation marks (for example, "myjnlDB..myjnlTABLE").
|
| Tip | If JOURNAL_TABLE is not specified, a journal is not created. If JOURNAL_TABLE is specified and a journal table exists, the journal is appended. |
specifies one or more columns by which to partition the input.
| Default | No partitioning occurs |
| Requirement | If the column name is case sensitive, you must use double
quotation marks(for example, "Profit", "Sales",
"Margin").
|
specifies one or more columns by which to order the input.
| Default | No reordering occurs |
| Requirement | If the column name is case sensitive, you must use double
quotation marks (for example, "Profit", "Sales",
"Margin").
|
'INPUT_WHERE=where name like ''%Jones%'''
modtable and the
model name is super. nzsql database username password select modelname, modelsequence from modtable where modelname like '%super%'
super.proc sql noerrorstop; connect to netezza (user=username password=xxxx server=myserver); select * from connection to netezza (select modelname, modelsequence from sasmodeltablename where modelname like '%super%'); disconnect netezza; quit;