Datafile line qualifiers

Purpose

Qualifiers modify the operation of ASReml. There are a large number of qualifiers, most are relevant to particular circumstances.
Qualifiers generally need to be specified in the correct context. Conseqently, 'data line' qualifiers must be on the line that specifies the data file, or on immediately following lines but before the model line. These are the only qualifiers that may be put on a separate line.

Definition

A qualifier is recognised as a letter sequence preceded by an explanation mark. Generally, the first three letters are sufficient to specify a qualifier although using the full name will normally make the job clearer to understand.

Some qualifiers have no argument. Most have a default argument. A few have multiple arguments.
An argument of '0' (zero) is normally equivalent to not specifying the qualifier.

Location

These datafile line qualifiers are placed either on the line that specifies the name of the data file, or on immediately following lines. A few, particularly those with multiple arguments like !DEFINE, !MBF, !PVAL, !SPLINE, !SUBSET must be put on separate lines by themselves.

Common qualifiers

!SKIP n causes ASReml to ignore the first n lines of the data file (assuming they are headings only)

Other qualifiers

!CSV forces COMMA or TAB to be used as field delimiters when reading the data file. Consecutive delimiters imply a missing value.

!FILTER v [!SELECT n | !EXCLUDE n ]
!SELECT n includes data records with the value n in the data to be analysed
!EXCLUDE n excludes data records with the value n in the data to be analysed

!FOLDER s sets the folder which holds the data file (when it differs from the current working folder)

!RREC m [!RSKIP n~s ] } reads the data from up to m data records after the nth instance of string s.

!DATAFILE f resets the data file name.

!FORMAT s provides a Fortran like format statement for reading the data.

!READ n sets the number of data fields to read when there are extra data fields that need to be read.

!RECODE indicates that the pedigree variables in the binary data file are not alredy coded according to the pedigree file.

!SPECIALCHAR cause # in the data file to be treated as a normal character; otherwise it has its special meaning of defining a comment to be stripped out.

Comma/TAB separated data.

Comma and TAB delimited files may have empty cells which are treated as missing values. A line beginning with TAB/Comma implies a leading missing value. Consecutive Comma/TAB characters imply a missing value. Comma separated files are recognised either from the file suffix .csv or specification of !CSV. TAB delimited files are recognised if the !CSV qualifier is set.

Spatial analysis of a multi-environment trial

When !SECTION site !ROWFAC row !COLFAC column
was specified in ASReml 3, it generated the R-structure lines to fit an AR1 x AR1 variance structure for each site. But that residual structure is now easily specified with the new syntax:

 residual at(site).ar1(row).ar1(col)

In ASReml 4, these 3 qualifiers will cause ASReml to 'complete the grid' provided 'mv' is included in the model. That is, after the data has been read in, any missing site/row/column grid cells (combinations) are created as extra data records. The residual statement causes them to be sorted into the appropriate order corresponding to the R structure specification.
Read More.