dsv_tsv_csv.format.txt

               
   DSV TSV CSV  
               


ALTERNATIVES ==>                  #See serialization formats summary

STANDARDS ==>                     #  - TSV: IANA MIME type
                                  #  - CSV: RFCs 4180, 7111

GOAL ==>                          #Simple way for an ARR|OBJ_ARR:
                                  #  - to persist it
                                  #  - to edit it with text editor
                                  #  - to parse it (good interoperability)
                                  #     - databases/spreadsheet/statistics programs can often open it
                                  #     - many Unix commands support delimiters: cut, paste, join, sort, uniq, awk
                                  #No size limits

DSV ==>                           #"Delimiter-separated values"
                                  #Encode an OBJ_ARR (if header) or ARR_ARR (if not header):
                                  #  - first dimension (records):
                                  #     - newline-separated
                                  #     - first record is header (optional)
                                  #     - blank records are ignored
                                  #  - second dimension (fields):
                                  #     - separated by any given character, e.g. , Tab Space | or :
                                  #     - escape delimiter/newline/" by encloding whole value with "..."
                                  #        - escape " by doubling it: ""
                                  #     - each record must have same number of fields
                                  #Can be any charset

TSV ==>                           #"Tab-separated values"
                                  #DSV using Tab as delimiter
                                  #Extension: .tsv or .tab
                                  #MIME type: text/tab-separated-values
                                  #No escaping: tabs are disallowed in field values

CSV ==>                           #"Comma-separated values"
                                  #DSV using , as delimiter
                                  #Extension: .csv
                                  #MIME type:
                                  #  - text/csv
                                  #  - parameters:
                                  #     - header "present|absent"
                                  #     - charset (def: ASCII)
                                  #Newline is CR+LF (last one optional)
                                  #Can use #FRAGMENT to select rows/cols/cells:
                                  #  - #row|col=NUMM[-NUMM2];...
                                  #  - #cell=ROW_NUMM,COL_NUMM[-ROW_NUMM2,COL_NUMM2];...
                                  #     - NUMM: NUM|*
                                  #        - NUM starts at 1
                                  #        - * means last
                                  #     - NUMM2 must be > NUMM
                                  #     - errors or non-existing ranges are ignored

VARIATIONS ==>                    #Implementations can sometimes be loose and not follow those standards.
                                  #Possible variations:
                                  #  - delimiter
                                  #  - newline
                                  #  - mandatory|optional header
                                  #  - blank records: ignored or not
                                  #  - skipping last fields if empty
                                  #  - escaping
                                  #  - charset