Node ‐ Plate Row Converter

Description

Currently this node expects any type of input columns (string, numeric (integer, double), but also others, which does not make sense) containing row positions. This node should expect string input columns and numeric (integer) columns containing row positions.

Expected row position format:

• row position as string may be one or two letters (upper and lower case)
• row position as numeric should be integers
• supports plate formats up to 1536 well plates

Examples: E, c, AE, 3, 12, 37

New columns will be created containing the plate row identifier (in case of input is string) and plate row letters (in case of numeric data) respectively. Currently the node can do both in one go, for each selected input it determins the type and applies the corresponding conversion. Double columns in some cases also work, but only first digits (numbers before the .) are considered.

DEV: Row pattern to match (string)([a-zA-Z]{1,2})

Icon

Category

/community/hcstools/utilities

Ports

• Input table: Table with column(s) containing row position to convert
• Output table: Table with (additional) column(s) of plate row identifier/letters

Options

• Shall the GUI follow same appearance/functionality as Expand Well Position '(a)' or should it be possible to select multiple columns like it is implemented now '(b)'?
  • => see Open Questions
  • (a) Row Position: Choose the column containing the row position to split
    DEV: The first string column matching .*[Rr]ow.* will be preselected, the first string column otherwise
  • (b) Row Position: Choose the multiple columns containing the row position to split
• Delete source column(s): If checked, the row position column(s) will be removed from the table
• Position of output column(s): New column(s) can be either appended to the end of the table or inserted after the row position column(s) (default)
  • End of the table: Appends the new column(s)
  • (Default) Behind row position column(s): New column(s) will be placed next to the row position column(s) (for multiple columns maybe tricky to implement!?)
• Output columns New column(s) may get default name(s) or custom name(s) may be set (for multiple columns maybe tricky to implement!?)
  • (Default) Keep default column name: Output columns will have same name as input column
    DEV: If column names already present, new columns will get typical KNIME suffix like (#1)
  • Rename: Give custom name(s) to outputput column(s)
    • (Optional) Column name (plate row identifier) Choose a name for the column containing the plate row identifier

Execute

• Missing Value Handling: will result in missing values (no warning)
• Cell content does not match well pattern: see Open Questions

Open Questions / Issues

• General GUI and function:

  1. MS: Shall the GUI follow same appearance/functionality as Expand Well Position (a) or ....
  2. MS: ... should it be possible to select multiple columns like it is implemented now (b)?
  3. MS: Should the node be able to convert doubles? I would say that is not neccessary and a bit odd implementation!?

• Cell content does not match well pattern:

  1. AJ: do we want an option 'fail on error' which (if selected) makes the node fail?
  • MS: fine with that, but if too complex, without is fine as well. Default is 'fail on error', then upon failing, this can be set to 'false'
  2. AJ: if not failing it would be nice to create a warning summary reporting on failing content. output would then be missing values. At the moment this is not possible.
  • MS: agree, but id not possible for now we have to live with that

• Category
  AJ: I don't like the category util. Shall we rename it? (at least to Utilities)

  • MS: totally agree