I've modified the CSV sample from Expert F# to my needs. I don't wann be forced to use the csv schema as defined by column rows. Therefore I've done two major modifications. 1. remove the permutation 2. added a new column name option to the ColumnAttribute 3. added a name to csv index mapping So basically you now have 3 options. 1. Don't annotate your record at all and use it as POCO. The order of the record fields is mapped directly to the order in the csv. UPDATE: I don't recommend this any more. As of the writing of this snippet I wasn't aware of the fact, that field order isn't guaranted by the reflection mechanism. 2. Use the index option of the ColumnAttribute. Same as before. 3. Use the name option. This is what I've looked for. I've to deal with tons of csv that has more columns I'm interested in. Have a look at the sample usage below. I've moved the type conversion out of the CsvReader class in order to be easyly expandable with custom type conversation (i.e. for combined column values - denormalized data)
4 people like thisPosted: 13 years ago by Rainer Schuster
High-performance JSON parsing with System.Text.Json and F#
1 people like thisPosted: 11 months ago by Tuomas Hietanen
This is a reader monad, whit the difference that multiple reads access different parts of the state. This is done by building up a tuple (in the bind operator) that represents different parts of the state that can be accessed by different parts of the computation.
5 people like thisPosted: 8 years ago by Tomas Petricek