Two-way table - serial
Requirements for this data format are:
- The data file should have 4 columns, representing 1) tester (e.g., test environment), 2) block (replicate), 3) entry (e.g., genotype), and 4) measured value for the trait, respectively.
- The first row is the header, containing the names of the columns, 1) "Tester name", 2) "Block name", 3) "Entry name", and 4) "Trait name" four items. You could put "End" after the trait name, but this is not required.
- Each subsequent row contains a single observation (a data point), and the 4 columns are for 1) the tester name, 2) the block (replication) name, 3) the entry name, and 4) the measured value, respectively.
- The data do not have to be balanced in terms of entry by tester combinations. The number of blocks can differ for different testers.
- There is no need to indicate missing values. If you do, missing cell can be indicated by anything except that it cannot be left empty for text files.
- The number of missing cells, if any, will be replaced by the respective tester means and the user will be notified once missing cells are detected.
See the following for an example:
Note: This format is a special case of the Three-way data - parallel format in which there is only one trait.