The SELECT statement enables you to specify the parts of speech or entities or attributes that you want to include in or exclude from your analysis. Exclusion by the SELECT statement is different from exclusion that is indicated by the _keep variable in the OUTTERMS= data table. Terms that are excluded by the SELECT statement cannot be included in the OUTTERMS= data table, whereas terms that have a _keep value of N can be included in the OUTTERMS= data table if the SHOWDROPPEDTERMS option is specified. Terms that are excluded by the SELECT statement are excluded from the OUTPOS= data table, but terms that have a _keep value of N are included in the OUTPOS= data table. Table 3 summarizes the options that you can specify in this statement. The options are then described fully in syntactic order.
Table 3: SELECT Statement Options
| Option | Description |
|---|
|
label-list | Specifies one or more labels of terms that are to be ignored or kept in your analysis |
|
GROUP= | Specifies whether the labels are parts of speech, entities, or attributes |
|
IGNORE | Ignores terms whose labels are specified in the label-list
|
|
KEEP | Keeps terms whose labels are specified in the label-list |
You must specify a label-list and either the IGNORE or KEEP option, as follows:
-
label-list
specifies one or more labels that are parts of speech, entities, or attributes. Each label must be enclosed in double quotation marks and separated by spaces from other labels. Labels are not case-sensitive. Terms that have these labels are either ignored during parsing (when the IGNORE option is specified) or kept in the parsing results in the OUTPOS= and OUTTERMS= data tables (when the KEEP option is specified). Table 6 shows all possible part-of-speech tags. Table 7 shows all valid English entities. The attribute variable in Table 12 shows all possible attributes.
-
IGNORE
ignores during parsing all terms whose labels are specified in the label-list, but keeps all other terms in the parsing results (in the OUTPOS= and OUTTERMS= data tables).
-
KEEP
keeps in the parsing results (in the OUTPOS= and OUTTERMS= data tables) only the terms whose labels are specified in the label-list.
You can also specify the following option:
-
GROUP="ATTRIBUTES" | "ENTITIES" | "POS"
specifies whether the labels are attributes, entities, or parts of speech. The group type must be enclosed in double quotation marks and is not case-sensitive. All labels that are specified in the label-list in the same SELECT statement should belong to the specified group. If you need to select labels from more than one group, you can use multiple SELECT statements (one for each group that you need to select from). You cannot specify multiple SELECT statements for the same group. By default, Num and Punct in the "ATTRIBUTES" group are ignored, but this default is overridden by a SELECT statement that specifies GROUP="ATTRIBUTES". By default, GROUP="POS".