Question

如果我有以下数据库结构：

TBL1

|id        | EYearMonth       | 
| 1        |     1617         |  
| 2        |     1618         |  
| 3        |     1619         |  
| 4        |     1620         |     
| 5        |     1621         |  
| 6        |     1622         | 
| 7        |     1623         | 
| 8        |     1624         | 
| 9        |     1625         | 
| 10       |     1626         | 
| 11       |     1627         | 
| 12       |     1628         |

TBL2

|id        | Value            | Serial#
| 1        |     1617         | 1068 
| 2        |     1618         | 1104
| 3        |     1624         | 1215

我真正想要的是以下内容：

结果

|id        | EYearMonth       | Serial#
| 1        |     1617         | 1068
| 2        |     1618         | 1104
| 3        |     1619         | 1104
| 4        |     1620         | 1104  
| 5        |     1621         | 1104  
| 6        |     1622         | 1104  
| 7        |     1623         | 1104  
| 8        |     1624         | 1215
| 9        |     1625         | 1215
| 10       |     1626         | 1215
| 11       |     1627         | 1215
| 12       |     1628         | 1215

如何制作此结果？请帮助我

Answer 1

您可以使用CROSS APPLY和TOP：

SELECT *
FROM tbl1 t1
CROSS APPLY(
    SELECT TOP 1 t2.[Serial#]
    FROM tbl2 t2
    WHERE t2.Value <= t1.EYearMonth
    ORDER BY t2.Value DESC
)t2

ONLINE DEMO

Answer 2

以下查询将起作用：

tvars = tf.trainable_variables()
grads = tf.gradients(cost, tvars)
gradients = zip(grads, tvars)
# The following block plots for every trainable variable
#  - Histogram of the entries of the Tensor
#  - Histogram of the gradient over the Tensor
#  - Histogram of the grradient-norm over the Tensor
for gradient, variable in gradients:
  if isinstance(gradient, ops.IndexedSlices):
    grad_values = gradient.values
  else:
    grad_values = gradient

  h1 = tf.histogram_summary(variable.name, variable)
  h2 = tf.histogram_summary(variable.name + "/gradients", grad_values)
  h3 = tf.histogram_summary(variable.name + "/gradient_norm",

sql join最近大于Value

2 个答案: