Skip to content

JSONLine

Bases: ReadWriteFileFormat

JSONLine file format (each line of file contains a JSON object). support hooks

Based on Spark JSON file format.

Supports reading/writing files with .json extension with content like:

example.json
{"key": "value1"}
{"key": "value2"}

Added in 0.9.0

Examples

Note

You can pass any option mentioned in official documentation. Option names should be in camelCase!

The set of supported options depends on Spark version.

from onetl.file.format import JSONLine

jsonline = JSONLine(encoding="UTF-8", mode="PERMISSIVE")

Warning

Written files have extension .json, not .jsonl or .jsonline.

from onetl.file.format import JSONLine

jsonline = JSONLine(encoding="UTF-8", compression="gzip")
Source code in onetl/file/format/jsonline.py 
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
@support_hooks
class JSONLine(ReadWriteFileFormat):
    """
    JSONLine file format (each line of file contains a JSON object). [![support hooks](https://img.shields.io/badge/%20-support%20hooks-blue)](/hooks/)

    Based on [Spark JSON](https://spark.apache.org/docs/latest/sql-data-sources-json.html) file format.

    Supports reading/writing files with `.json` extension with content like:

    ```json title="example.json"
    {"key": "value1"}
    {"key": "value2"}
    ```
    !!! success "Added in 0.9.0"

    Examples
    --------

    !!! note

        You can pass any option mentioned in
        [official documentation](https://spark.apache.org/docs/latest/sql-data-sources-json.html).
        **Option names should be in** `camelCase`!

        The set of supported options depends on Spark version.

    === "Reading files"
        ```python
        from onetl.file.format import JSONLine

        jsonline = JSONLine(encoding="UTF-8", mode="PERMISSIVE")
        ```
    === "Writing files"

        !!! warning

            Written files have extension `.json`, not `.jsonl` or `.jsonline`.

        ```python
        from onetl.file.format import JSONLine

        jsonline = JSONLine(encoding="UTF-8", compression="gzip")
        ```
    """

    name: ClassVar[str] = "json"

    multiLine: Literal[False] = False

    encoding: Optional[str] = None
    """
    Encoding of the JSONLine files.
    Default `UTF-8`.
    """

    lineSep: Optional[str] = None
    """
    Character used to separate lines in the JSONLine files.

    Defaults:
      * Try to detect for reading (`\\r\\n`, `\\r`, `\\n`)
      * `\\n` for writing.
    """

    compression: Union[str, Literal["none", "bzip2", "gzip", "lz4", "snappy", "deflate"], None] = None
    """
    Compression codec of the JSONLine file.
    Default `none`.

    !!! note

        Used only for writing files.
    """

    ignoreNullFields: Optional[bool] = None
    """
    If `True` and field value is `null`, don't add field into resulting object
    Default is value of `spark.sql.jsonGenerator.ignoreNullFields` (`True`).

    !!! note

        Used only for writing files.
    """

    allowComments: Optional[bool] = None
    """
    If `True`, add support for C/C++/Java style comments (`//`, `/* */`).
    Default `False`, meaning that JSONLine files should not contain comments.

    !!! note

        Used only for reading files.
    """

    allowUnquotedFieldNames: Optional[bool] = None
    """
    If `True`, allow JSON object field names without quotes (JavaScript-style).
    Default `False`.

    !!! note

        Used only for reading files.
    """

    allowSingleQuotes: Optional[bool] = None
    """
    If `True`, allow JSON object field names to be wrapped with single quotes (`'`).
    Default `True`.

    !!! note

        Used only for reading files.
    """

    allowNumericLeadingZeros: Optional[bool] = None
    """
    If `True`, allow leading zeros in numbers (e.g. `00012`).
    Default `False`.

    !!! note

        Used only for reading files.
    """

    allowNonNumericNumbers: Optional[bool] = None
    """
    If `True`, allow numbers to contain non-numeric characters, like:
      * scientific notation (e.g. `12e10`).
      * positive infinity floating point value (`Infinity`, `+Infinity`, `+INF`).
      * negative infinity floating point value (`-Infinity`, `-INF`).
      * Not-a-Number floating point value (`NaN`).

    Default `True`.

    !!! note

        Used only for reading files.
    """

    allowBackslashEscapingAnyCharacter: Optional[bool] = None
    """
    If `True`, prefix `\\` can escape any character.
    Default `False`.

    !!! note

        Used only for reading files.
    """

    allowUnquotedControlChars: Optional[bool] = None
    """
    If `True`, allow unquoted control characters (ASCII values 0-31) in strings without escaping them with `\\`.
    Default `False`.

    !!! note

        Used only for reading files.
    """

    mode: Optional[Literal["PERMISSIVE", "DROPMALFORMED", "FAILFAST"]] = None
    """
    How to handle parsing errors:
      * `PERMISSIVE` - set field value as `null`, move raw data to [columnNameOfCorruptRecord][] column.
      * `DROPMALFORMED` - skip the malformed row.
      * `FAILFAST` - throw an error immediately.

    Default is `PERMISSIVE`.

    !!! note

        Used only for reading files.
    """

    columnNameOfCorruptRecord: Optional[str] = Field(default=None, min_length=1)
    """
    Name of column to put corrupt records in.
    Default is `_corrupt_record`.

    !!! warning

        If DataFrame schema is provided, this column should be added to schema explicitly:

        ```python
        from onetl.connection import SparkLocalFS
        from onetl.file import FileDFReader
        from onetl.file.format import JSONLine

        from pyspark.sql.types import StructType, StructField, TimestampType, StringType

        spark = ...

        schema = StructType(
            [
                StructField("my_field", TimestampType()),
                StructField("_corrupt_record", StringType()),  # <-- important
            ]
        )

        jsonline = JSONLine(mode="PERMISSIVE", columnNameOfCorruptRecord="_corrupt_record")

        reader = FileDFReader(
            connection=connection,
            format=jsonline,
            df_schema=schema,  # < ---
        )
        df = reader.run(["/some/file.jsonl"])
        ```
    !!! note

        Used only for reading files.
    """

    samplingRatio: Optional[float] = Field(default=None, ge=0, le=1)
    """
    While inferring schema, read the specified fraction of file rows.
    Default `1`.

    !!! note

        Used only for reading files.
    """

    primitivesAsString: Optional[bool] = None
    """
    If `True`, infer all primitive types (string, integer, float, boolean) as strings.
    Default `False`.

    !!! note

        Used only for reading files.
    """

    prefersDecimal: Optional[bool] = None
    """
    If `True`, infer all floating-point values as `Decimal`.
    Default `False`.

    !!! note

        Used only for reading files.
    """

    dropFieldIfAllNull: Optional[bool] = None
    """
    If `True` and inferred column is always null or empty array, exclude if from DataFrame schema.
    Default `False`.

    !!! note

        Used only for reading files.
    """

    dateFormat: Optional[str] = Field(default=None, min_length=1)
    """
    String format for `DateType()` representation.
    Default is `yyyy-MM-dd`.
    """

    timestampFormat: Optional[str] = Field(default=None, min_length=1)
    """
    String format for `TimestampType()` representation.
    Default is `yyyy-MM-dd'T'HH:mm:ss[.SSS][XXX]`.
    """

    timestampNTZFormat: Optional[str] = Field(default=None, min_length=1)
    """
    String format for `TimestampNTZType()` representation.
    Default is `yyyy-MM-dd'T'HH:mm:ss[.SSS]`.

    !!! note

        Added in Spark 3.2.0
    """

    timezone: Optional[str] = Field(default=None, min_length=1, alias="timeZone")
    """
    Allows to override timezone used for parsing or serializing date and timestamp values.
    By default, `spark.sql.session.timeZone` is used.
    """

    locale: Optional[str] = Field(default=None, min_length=1)
    """
    Locale name used to parse dates and timestamps.
    Default is `en-US`.

    !!! note

        Used only for reading files.
    """

    class Config:
        known_options: frozenset[str] = frozenset()
        extra = "allow"

    @slot
    def check_if_supported(self, spark: SparkSession) -> None:
        # always available
        pass

    def __repr__(self):
        options_dict = self.dict(by_alias=True, exclude_none=True, exclude={"multiLine"})
        options_dict = dict(sorted(options_dict.items()))
        options_kwargs = ", ".join(f"{k}={v!r}" for k, v in options_dict.items())
        return f"{self.__class__.__name__}({options_kwargs})"

allowBackslashEscapingAnyCharacter = None class-attribute instance-attribute

If True, prefix \ can escape any character. Default False.

Note

Used only for reading files.

allowComments = None class-attribute instance-attribute

If True, add support for C/C++/Java style comments (//, /* */). Default False, meaning that JSONLine files should not contain comments.

Note

Used only for reading files.

allowNonNumericNumbers = None class-attribute instance-attribute

If True, allow numbers to contain non-numeric characters, like: * scientific notation (e.g. 12e10). * positive infinity floating point value (Infinity, +Infinity, +INF). * negative infinity floating point value (-Infinity, -INF). * Not-a-Number floating point value (NaN).

Default True.

Note

Used only for reading files.

allowNumericLeadingZeros = None class-attribute instance-attribute

If True, allow leading zeros in numbers (e.g. 00012). Default False.

Note

Used only for reading files.

allowSingleQuotes = None class-attribute instance-attribute

If True, allow JSON object field names to be wrapped with single quotes ('). Default True.

Note

Used only for reading files.

allowUnquotedControlChars = None class-attribute instance-attribute

If True, allow unquoted control characters (ASCII values 0-31) in strings without escaping them with \. Default False.

Note

Used only for reading files.

allowUnquotedFieldNames = None class-attribute instance-attribute

If True, allow JSON object field names without quotes (JavaScript-style). Default False.

Note

Used only for reading files.

columnNameOfCorruptRecord = Field(default=None, min_length=1) class-attribute instance-attribute

Name of column to put corrupt records in. Default is _corrupt_record.

Warning

If DataFrame schema is provided, this column should be added to schema explicitly:

from onetl.connection import SparkLocalFS
from onetl.file import FileDFReader
from onetl.file.format import JSONLine

from pyspark.sql.types import StructType, StructField, TimestampType, StringType

spark = ...

schema = StructType(
    [
        StructField("my_field", TimestampType()),
        StructField("_corrupt_record", StringType()),  # <-- important
    ]
)

jsonline = JSONLine(mode="PERMISSIVE", columnNameOfCorruptRecord="_corrupt_record")

reader = FileDFReader(
    connection=connection,
    format=jsonline,
    df_schema=schema,  # < ---
)
df = reader.run(["/some/file.jsonl"])

Note

Used only for reading files.

compression = None class-attribute instance-attribute

Compression codec of the JSONLine file. Default none.

Note

Used only for writing files.

dateFormat = Field(default=None, min_length=1) class-attribute instance-attribute

String format for DateType() representation. Default is yyyy-MM-dd.

dropFieldIfAllNull = None class-attribute instance-attribute

If True and inferred column is always null or empty array, exclude if from DataFrame schema. Default False.

Note

Used only for reading files.

encoding = None class-attribute instance-attribute

Encoding of the JSONLine files. Default UTF-8.

ignoreNullFields = None class-attribute instance-attribute

If True and field value is null, don't add field into resulting object Default is value of spark.sql.jsonGenerator.ignoreNullFields (True).

Note

Used only for writing files.

lineSep = None class-attribute instance-attribute

Character used to separate lines in the JSONLine files.

Defaults
  • Try to detect for reading (\r\n, \r, \n)
  • \n for writing.

locale = Field(default=None, min_length=1) class-attribute instance-attribute

Locale name used to parse dates and timestamps. Default is en-US.

Note

Used only for reading files.

mode = None class-attribute instance-attribute

How to handle parsing errors
  • PERMISSIVE - set field value as null, move raw data to [columnNameOfCorruptRecord][] column.
  • DROPMALFORMED - skip the malformed row.
  • FAILFAST - throw an error immediately.

Default is PERMISSIVE.

Note

Used only for reading files.

prefersDecimal = None class-attribute instance-attribute

If True, infer all floating-point values as Decimal. Default False.

Note

Used only for reading files.

primitivesAsString = None class-attribute instance-attribute

If True, infer all primitive types (string, integer, float, boolean) as strings. Default False.

Note

Used only for reading files.

samplingRatio = Field(default=None, ge=0, le=1) class-attribute instance-attribute

While inferring schema, read the specified fraction of file rows. Default 1.

Note

Used only for reading files.

timestampFormat = Field(default=None, min_length=1) class-attribute instance-attribute

String format for TimestampType() representation. Default is yyyy-MM-dd'T'HH:mm:ss[.SSS][XXX].

timestampNTZFormat = Field(default=None, min_length=1) class-attribute instance-attribute

String format for TimestampNTZType() representation. Default is yyyy-MM-dd'T'HH:mm:ss[.SSS].

Note

Added in Spark 3.2.0

timezone = Field(default=None, min_length=1, alias='timeZone') class-attribute instance-attribute

Allows to override timezone used for parsing or serializing date and timestamp values. By default, spark.sql.session.timeZone is used.