ultra-csv.core — ultra-csv 0.2.3

close! clj

(close! f)

Closes and cleans the io ressources used. Can be used on a greedy generator or any line of a lazy seqs of results

csv-line clj

(csv-line spec data)

csv-line-reader clj

(csv-line-reader {:keys [header? field-names schema guess-types? strict?
                         counter-step silent? limit nullable-fields? coercers]
                  :or {guess-types? true strict? true coercers {}}
                  :as opts})

This function returns a reader function that accepts inputs of one String line at a time. It takes the same options as read-csv minus some processing and the file and analysis options.

guess-spec clj

(guess-spec uri)

(guess-spec
  uri
  {:keys [header? field-names field-names-fn schema encoding guess-types?
          delimiter nullable-fields? keywordize-keys? sample-size bom skip]
   :or {guess-types? true nullable-fields? true sample-size 100 skip 0}
   :as opts})

This function takes a source of csv lines (either a Reader, InputStream or String URI) and tries to guess the specs necessary to parse it. You can use the option map to specify some values for the spec. Recognised options are:

Analysis options

• header?: Whether the file as a header on the first line
• sample-size: number of lines on which heuristics are applied. Defaults to 100
• guess-types?: Whether to try to guess types for each field. Defaults to true
• skip: Number of lines to skip before starting to parse. Defaults to 0

File options

• encoding: Character encoding for the file

Processing options

• field-names-fn: fn to apply to the name of each field. Defaults to trim function
• nullable-fields?: Whether the values are optionals. Defaults to true
• keywordize-keys?: Whether to turn fields into keywords. Defaults to true

Format options

• delimiter: Character used as a delimiter
• schema: Schema to validate and coerce output
• field-names: Names for the csv fields, in order

load-spec-str clj

(load-spec-str spec-string)

Loads a spec structure from String form. BEWARE The uses eval on the passed String, only use on inputs that are ABSOLUTELY SAFE

make-date-coercer clj

(make-date-coercer fmt)

(make-date-coercer fmt timezone)

Makes a date coercer to use with the coercer option of read-csv. Given a formatter String and an optional timezonr String, this will return a ready to use function. Details on the format can be found at http://docs.oracle.com/javase/7/docs/api/java/text/SimpleDateFormat.html

read-csv clj

(read-csv uri)

(read-csv uri
          {:keys [header? field-names field-names-fn schema encoding
                  guess-types? strict? greedy? counter-step silent? limit
                  skip-analysis? nullable-fields? keywordize-keys? coercers bom
                  skip]
           :or {guess-types? true
                strict? true
                nullable-fields? true
                keywordize-keys? true
                coercers {}
                skip 0}
           :as opts})

This function takes a source of csv lines (either a Reader, InputStream or String URI) and returns the parsed results. If headers are found on the file or field-names where given as options, it will return a collection of one map per line, associating each field name with its value. If not, one vector will be returned for each line, in order.

You can use the option map to specify some values for the spec. Recognised options are:

Analysis options

• header?: Whether the file as a header on the first line
• sample-size: number of lines on which heuristics are applied. Defaults to 100
• guess-types?: Whether to try to guess types for each field. Defaults to true
• skip-analysis?: Whether to completely bypass analysis and only use spec Defaults to false
• skip: Number of lines to skip before starting to parse. Defaults to 0

Processing options

• greedy?: If true returns a function that can be used as a generator, returning one line with each call, else returns a lazy seq of lines. Defaults to false
• strict?: Whether to throw exception on reading of validation error, or just skip it. Defaults to true
• silent?: Whether there should be error messages emitted on stderr when skipping Exceptions. Defaults to false
• limit: Closes and cleans the io ressources after reading this many lines. Useful for sampling
• field-names-fn: fn to apply to the name of each field. Can be used to sanitize header names. Defaults to trim function
• nullable-fields?: Whether the values are optionals. Defaults to true
• keywordize-keys?: Whether to turn fields into keywords. Defaults to true

Format options

• delimiter: Character used as a delimiter

• schema: Schema to validate and coerce output

• coercers: A map associating a type as key with a function from String to that type. For example:

  {:coercers {java.util.Date (make-date-coercer "yyyyMMdd")}}
  

• field-names: Names for the csv fields, in order

write-csv! clj

(write-csv! uri data)

(write-csv! uri
            {:keys [bom encoding field-names coercers header? append]
             :as spec
             :or {bom :none header? true}}
            data)

This function takes a target (either a Reader, InputStream or String URI), a spec of options and a seq of either maps or seqs. Each data structure is written to the target csv as a single line.

Right now, custom encoders are ignored and each value is just coerced to its String format.

The option map is the same as read-csv with these added options:

Write options

• append: if set to true, opens the given uri in append mode. header? option is then ignored.

write-spec clj

(write-spec spec)

Writes a spec struct to a String