0
# Data I/O Operations
1

2
High-level functions for saving and loading zarr data structures to and from storage. These provide convenient interfaces for persistence operations with support for various data formats and storage backends.
3

4
## Capabilities
5

6
### Data Loading
7

8
```python { .api }
9
def load(
10
    store: StoreLike,
11
    path: str = None,
12
    **kwargs
13
) -> Any
14
```
15

16
Load zarr array data into memory as a numpy array or nested structure.
17

18
**Parameters:**
19
- `store`: Storage location containing zarr data
20
- `path`: Path within store to load from
21

22
**Returns:** Numpy array for single arrays, dict-like structure for groups
23

24
### Single Array Saving
25

26
```python { .api }
27
def save_array(
28
    store: StoreLike,
29
    arr: ArrayLike, 
30
    path: str = None,
31
    **kwargs
32
) -> None
33
```
34

35
Save a single array to zarr format.
36

37
**Parameters:**
38
- `store`: Storage location to save to
39
- `arr`: Array data to save
40
- `path`: Path within store to save to
41

42
### Group Saving
43

44
```python { .api }
45
def save_group(
46
    store: StoreLike,
47
    *args,
48
    path: str = None,
49
    **kwargs
50
) -> None
51
```
52

53
Save a group structure with multiple arrays.
54

55
### General Save Function
56

57
```python { .api }
58
def save(
59
    file: Union[str, StoreLike],
60
    *args,
61
    **kwargs
62
) -> None
63
```
64

65
General-purpose save function that handles both arrays and groups.
66

67
**Parameters:**
68
- `file`: File path or storage location
69
- `*args`: Arrays or array-like data to save
70
- `**kwargs`: Additional arrays as keyword arguments
71

72
## Usage Examples
73

74
### Basic Array I/O
75

76
```python
77
import zarr
78
import numpy as np
79

80
# Create and save array
81
data = np.random.random((1000, 1000))
82
zarr.save('dataset.zarr', data)
83

84
# Load array back
85
loaded_data = zarr.load('dataset.zarr')
86
```
87

88
### Multiple Array I/O
89

90
```python
91
# Save multiple arrays
92
arr1 = np.random.random((500, 500))
93
arr2 = np.random.random((300, 300))
94

95
zarr.save('multi_arrays.zarr', 
96
          temperature=arr1, 
97
          humidity=arr2)
98

99
# Load returns dict-like structure
100
data = zarr.load('multi_arrays.zarr')
101
temp = data['temperature']
102
humidity = data['humidity']
103
```
104

105
### Advanced I/O with Compression
106

107
```python
108
from zarr.codecs import BloscCodec
109

110
# Save with compression
111
zarr.save('compressed.zarr', 
112
          data, 
113
          compressor=BloscCodec(cname='zstd', clevel=3))
114

115
# Load automatically handles decompression
116
loaded = zarr.load('compressed.zarr')
117
```