column - project columns from a starbase data table

SYNOPSYS

column  [-i input] [-o output] [options] [column] ...
project [-i input] [-o output] [options] [column] ...

column is used to create, delete, format and choose columns in a table.

column reads a starbase data table from its standard input and prints a table with the columns given as arguments on its standard output. If a given column does not exist it is created in the output. If no column are given as arguments all the columns in the table are output.

column and project are the same program.

OPTIONS

• -i file Read input from file instead of the standard input.
• -o file Write output to file instead of the standard output.
• -t template Only the tables whose table names (first line in the table header) match template are output. Care should be taken with this option, unnamed tables are skipped if it is specified. This is the primary mechanism for selecting a table from a multi-table file.

• -a Output all the columns in the table. This is useful when adding columns to a table. This argumant is positional. All the columns in the input table that have not yet appeared on the command line are inserted here.

  Here columns names "Result" and "Comment" are appended to the right side of a table:

          john@panic: column -a Result Comment
          X       Y       Result  Comment
          -       -       ------  -------
          100     10
  

  Here columns are inserted at the left side of a table:

          john@panic: column -a Result Comment | justify
          Result  Comment X       Y
          ------  ------- ---     --
                          100     10
  

• -b Print only the body (rows) of the table.
• -hv Print out only the header values.
• -hV Print out only the header values as a value table.
• -hl Print out only the headline.
• -hd Print out only the dashline.
• -hh Print out only the rdb header - not the starbase values extension.
• -k Check the validity of the input table. Exit with status is 1 if the table is not ok.
• -u Read the standard input in unbuffered mode. This is a very useful option for taking the header part off of a starbase table and passing the rest to another unix program filter. See the source of headtable as an example.
• -v Verbose error checking. Print messages about offending rows to the standard output. If check is being run the entire table is checked. If only -k is specified the program exits when it finds the first error.
• -2 Print each row in the table twice. This option is useful with sortable to eliminate or select duplicate rows. This is the underlying mechanism of diffrtable and intertable.
• column=newcol rename a column. Columns may be renamed by any of the column programs. Before the rename option the column must be referenced by its old name, after by its new one. Renaming a column does not select it for output.

The use of any justification flags implies that all the columns will be justified. Justification options that are listed on the command line apply to all following columns until a different option is given. Options that are suffixed to a column name with ":" only apply to that column. The default justification is left for text and right for numbers. The output width of a column may be controlled by appending a width for the column in characters to a column name with a ",". A column can be reformatted by appending a sprintf format spec to the columns name with a "%".

• -d Justify the column with its default justification. Numeric values are right justified and text is left justified.
• -l Left justify.
• -c Center justify.
• -r Right justify.
• -e Expand the column.
• -w width

  Set the output column width. Care should be taken with the width option, column data will be truncated if they are width than the width specification.

• -s Skip the data in the column.
• -x Delete the column.
• -p Pass the column through with no justification.

Some examples of justification options with justify. Other column programs also support these options:

Justify the table:

    john@panic : justify < foo
    X       Name            Y      
    ------  --------        -------
       4.5  Star1           5000   
     435    Star2_bk           4.3 
    2300    Star3              2.04

Justify all the column, override the default left for text:

    john@panic : justify < foo -a -r Name
        Name        X       Y      
    --------        ------  -------
       Star1           4.5  5000   
    Star2_bk         435       4.3 
       Star3        2300       2.04

Do the same thing with an appended column option:

    john@panic : justify < foo -a Name:r
        Name        X       Y
    --------        ------  -------
       Star1           4.5  5000
    Star2_bk         435       4.3
       Star3        2300       2.04

Sepcify a column width for the name column:

    john@panic : justify Name,20 < foo
    Name                
    --------------------
    Star1               
    Star2_bk            
    Star3               

Column will accept column formatting options to reformat the ASCII representation of numeric columns. These options are invoked by following the column name with a printf style % specification.

Format some columns:

 john@panic : column X%7.2f Y%7.2f RA%12.3@ Dec%11.2@
 X       Y       RA      Dec
 -       -       --      ---
 512.00  512.00           12:30:34.501           34:00:54.45

Ask for all the column from the standard input

    john@panic: column <  input
    X       Y
    -       -
    0       1
    0       1

Ask for all the columns from a file specified with the -i switch put the output in a file specified with the -o switch.

    john@panic: column -i input -o output

Ask for the X column

    john@panic: column X < input
    X
    -
    0
    0

Number the rows

    john@panic: column -n < input
    Number  X       Y
    ------  -       -
    1       0       1
    2       0       1
    3       0       1

Rename a column.

    john@panic: column X=K < input
    K       Y
    -       -
    0       1
    0       1
    0       1

Output only the header without the rows

    john@panic: $ column -h < input
    Value Table

    X1      1
    Y1      1       2       3

    X       Y
    -       -

Output the rows wo/ the header.

    john@panic: $ column -b < input
    0       1
    1       2

Output the extended header part.

    john@panic: column -hv < input
    Value Table

    X1      1
    Y1      1       2       3

Output the header values as a value table

    john@panic: $ column -hV < val $
    Name    Value
    ----    -----
    X1      1
    Y1      1       2       3

Output the headline or the dashline.

    john@panic: $ column -hl < input
    X       Y
    john@panic: $ column -hd < input
    -       -

Add a column to the table using the -a all/add option.

    john@panic: $ column -a Z < input
    Z       X       Y
    -       -       -
            0       1
            0       1
            0       1

Justify one column with the default justifications pass the others through.

    john@panic: $ column AAAAAA:d B C D E < U.tab $
    AAAAAA  B       C       D       E
    ------  -       -       -       -
    1.00    112.00  3       4.00    5
    1      1113    4.00    5.0     6
    1      1113    4.00    5.0     6

Justify all column by default but pass one column through using the column :p flag.

    john@panic: $ column -d AAAAAA B C:p D E < U.tab $
    AAAAAA  B       C       D       E
    ------  ------- -       ------- -
      1.00   112.00 3          4.00 5
      1     1113    4.00       5.0  6

Columns can be selected by explicit name or by pattern matching. This can be effectivly used to select columns that have similar prefix or suffix charateers. Column name patterns use the same file selection meta chatacters that are used by the shell (*?[]). When selecting column for renaming (with the = option) the "matched" part of the column name can be used in the new name with the '?' character.

Examples

After a search -j operation, select all of the column from the keytable:

    john@panic: column '*_1' < search_results

On the same table rename the columns back to thier original names:

    john@panic: column '*_1=?' < search_results

• column, project project columns from a starbase data table.
• justify justify columns of a starbase data table.
• trim trim columns of an starbase data table.
• check check the format of a starbase data table.
• header print the header of a table file.
• headoff print the body (data rows) of a starbase data table file wo/ the header.
• headline print the column headline of a starbase data table.
• dashline print the dashline from a starbase data table.

• starbase.1 starbase data table introduction.
• starbase.3 starbase data table C library.
• starbase.5 starbase data table file format.