Changes between Version 52 and Version 53 of BluePrint/Importer


Ignore:
Timestamp:
01/23/11 08:54:22 (11 years ago)
Author:
Pat Tressel
Comment:

--

Legend:

Unmodified
Added
Removed
Modified
  • BluePrint/Importer

    v52 v53  
    126126== CSV import ==
    127127
    128 //Under Construction...falling...asleep...will continue when awake//
     128//Under Construction//
    129129
    130130This discusses only the back end of the CSV import, not the UI.
     
    226226Besides the observed forms, please propose others that we want to support:
    227227
    228 - (Comments please!)
     228- //(Comments please!)//
    229229
    230230==== Which should we implement in the initial version? ====
     
    235235- (Pat:) Single file, outer join, allow lists in cells.
    236236
    237 - (Comments please!)
     237- //(Comments please!)//
    238238
    239239=== Specifying the file layout and schema mapping ===
     
    302302- For concatenated tables, what is the table separator?
    303303
    304 - Is there a row with column names in the file, or are column names supplied
    305   separately?
     304- Are columns specified (in the schema mapping) by name or number?
     305 - If name, is there a row with column names in the file, or are column names
     306   supplied separately?
     307 - If number, is it 0-based or 1-based?
    306308
    307309- Column separator? (Popular separators other than commas are tabs, semicolons,
     
    322324==== Schema mapping specification: ====
    323325
     326Very broadly:
     327
     328- For non-object cells, specify: (Sheet, column), Simple, Xform:<xform> -> (Eden table, field)
     329- For list cells (i.e. where the list contents are each added as a
     330  separate record), specify: (Sheet, column), List, Xform:<xform> -> (Eden table, field)
     331- For object cells, specify: (Sheet, column), Object, Format:<format> (e.g. XML),
     332  { (tag hierarchy, i.e. tag1:tag2:tag3...), Xform:<xform> -> (Eden table, field), ... }
     333
     334Xform:<xform> is a transformation to be applied to the data. (E.g. if a name
     335in the primary table is to be used for its associated location, might want to
     336say "Location of <name>". Yes, I know that's not a good example. However, it
     337is one that has actually been done.)
     338
     339For the user specification, if not entered via a UI form, we could ask for a
     340common format like XML, or something more human-readable.
     341
     342- //(Suggestions for the user specification format?)//
     343
     344- //(Comments, especially, what cases have been omitted?)//
     345
    324346=== Implementation notes ===
    325347
     
    329351
    330352- Extract the data into XML.
     353
    331354- Call S3Resource import_xml().
    332355