0
# Dataset Management
1

2
Comprehensive dataset creation, management, and preparation for various ML tasks including tabular, image, text, video, and time series data. Vertex AI datasets provide managed data storage with automatic schema detection and data validation.
3

4
## Capabilities
5

6
### Tabular Datasets
7

8
Structured data management for classification, regression, and forecasting tasks with automatic schema detection and data quality analysis.
9

10
```python { .api }
11
class TabularDataset:
12
    @classmethod
13
    def create(
14
        cls,
15
        display_name: str,
16
        gcs_source: Union[str, Sequence[str]],
17
        bq_source: Optional[str] = None,
18
        project: Optional[str] = None,
19
        location: Optional[str] = None,
20
        labels: Optional[Dict[str, str]] = None,
21
        encryption_spec_key_name: Optional[str] = None,
22
        sync: bool = True,
23
        create_request_timeout: Optional[float] = None,
24
        **kwargs
25
    ) -> 'TabularDataset': ...
26

27
    def import_data(
28
        self,
29
        gcs_source: Optional[Union[str, Sequence[str]]] = None,
30
        bq_source: Optional[str] = None,
31
        import_schema_uri: Optional[str] = None,
32
        data_item_labels: Optional[Dict] = None,
33
        sync: bool = True,
34
        **kwargs
35
    ) -> 'TabularDataset': ...
36

37
    @property
38
    def column_names(self) -> List[str]: ...
39
    @property
40
    def schema(self) -> Dict[str, str]: ...
41
```
42

43
### Image Datasets
44

45
Image data management for classification, object detection, and segmentation tasks with support for various annotation formats.
46

47
```python { .api }
48
class ImageDataset:
49
    @classmethod
50
    def create(
51
        cls,
52
        display_name: str,
53
        gcs_source: str,
54
        import_schema_uri: str,
55
        data_item_labels: Optional[Dict] = None,
56
        project: Optional[str] = None,
57
        location: Optional[str] = None,
58
        labels: Optional[Dict[str, str]] = None,
59
        encryption_spec_key_name: Optional[str] = None,
60
        sync: bool = True,
61
        create_request_timeout: Optional[float] = None,
62
        **kwargs
63
    ) -> 'ImageDataset': ...
64

65
    def import_data(
66
        self,
67
        gcs_source: str,
68
        import_schema_uri: str,
69
        data_item_labels: Optional[Dict] = None,
70
        sync: bool = True,
71
        **kwargs
72
    ) -> 'ImageDataset': ...
73
```
74

75
### Text Datasets
76

77
Text data management for classification, entity extraction, and sentiment analysis with support for various text formats.
78

79
```python { .api }
80
class TextDataset:
81
    @classmethod
82
    def create(
83
        cls,
84
        display_name: str,
85
        gcs_source: Union[str, Sequence[str]],
86
        import_schema_uri: str,
87
        data_item_labels: Optional[Dict] = None,
88
        project: Optional[str] = None,
89
        location: Optional[str] = None,
90
        labels: Optional[Dict[str, str]] = None,
91
        encryption_spec_key_name: Optional[str] = None,
92
        sync: bool = True,
93
        create_request_timeout: Optional[float] = None,
94
        **kwargs
95
    ) -> 'TextDataset': ...
96
```
97

98
### Time Series Datasets
99

100
Specialized datasets for forecasting and time series analysis with support for multiple time series and hierarchical forecasting.
101

102
```python { .api }
103
class TimeSeriesDataset:
104
    @classmethod
105
    def create(
106
        cls,
107
        display_name: str,
108
        gcs_source: Union[str, Sequence[str]],
109
        bq_source: Optional[str] = None,
110
        project: Optional[str] = None,
111
        location: Optional[str] = None,
112
        labels: Optional[Dict[str, str]] = None,
113
        encryption_spec_key_name: Optional[str] = None,
114
        sync: bool = True,
115
        create_request_timeout: Optional[float] = None,
116
        **kwargs
117
    ) -> 'TimeSeriesDataset': ...
118
```
119

120
### Video Datasets
121

122
Video data management for action recognition, object tracking, and video classification tasks.
123

124
```python { .api }
125
class VideoDataset:
126
    @classmethod
127
    def create(
128
        cls,
129
        display_name: str,
130
        gcs_source: Union[str, Sequence[str]],
131
        import_schema_uri: str,
132
        data_item_labels: Optional[Dict] = None,
133
        project: Optional[str] = None,
134
        location: Optional[str] = None,
135
        labels: Optional[Dict[str, str]] = None,
136
        encryption_spec_key_name: Optional[str] = None,
137
        sync: bool = True,
138
        create_request_timeout: Optional[float] = None,
139
        **kwargs
140
    ) -> 'VideoDataset': ...
141
```
142

143
## Usage Examples
144

145
**Create tabular dataset:**
146
```python
147
import google.cloud.aiplatform as aiplatform
148

149
aiplatform.init(project='my-project', location='us-central1')
150

151
dataset = aiplatform.TabularDataset.create(
152
    display_name="customer-data",
153
    gcs_source="gs://my-bucket/customer_data.csv",
154
    labels={"purpose": "classification", "team": "ml"}
155
)
156

157
print(f"Dataset created: {dataset.resource_name}")
158
print(f"Column names: {dataset.column_names}")
159
```
160

161
**Create image dataset:**
162
```python
163
dataset = aiplatform.ImageDataset.create(
164
    display_name="product-images",
165
    gcs_source="gs://my-bucket/images/",
166
    import_schema_uri=aiplatform.schema.dataset.ioformat.image.single_label_classification
167
)
168
```