首先创建截图中的表格：

t1 = `T1`T2`T3`T4`T5`T6`T7
s1 = `A`A`A`A`A`A`
b1 = 0.35 0.34 0.33 0.28 0.29 0 0.3
a1 = 0.34 0.33 0.32 0.3 0 0 0.31
t = table(t1 as timestamp, s1 as symbol, b1 as bid, a1 as ask)

题主的问题，有以下两种方法可以参考：

方案一：ffill方法

select  *,ffill(iif(prev(bid)==0,NULL,prev(bid))) as last_bid, ffill(iif(prev(bid)==0,NULL,prev(timestamp))) as last_bid_timestamp,  ffill(iif(prev(ask)==0,NULL,prev(ask))) as last_ask,  ffill(iif(prev(ask)==0,NULL,prev(timestamp))) as last_ask_timestamp from t

方法二：使用cumlastNot函数

update t set last_bid = cumlastNot(bid, 0), last_ask = cumlastNot(ask, 0), last_bid_timestamp = cumlastNot(iif(prev(bid)!= 0, prev(timestamp), NULL), NULL), last_ask_timestamp = cumlastNot(iif(prev(ask)!= 0, prev(timestamp), NULL), NULL)

性能测试

数据量：构造100万行的bid，ask价格表

测试方案：通过timer函数统计单次计算的耗时

测试结果：

参考如下实现：

t1 = `T1`T2`T3`T4`T5`T6`T7
s1 = `A`A`A`A`A`A`
b1= 0.35 0.34 0.33 0.28 0.29 0 0.3
a1 =0.34 0.33 0.32 0.3 0 0 0.31
t = table(t1 as timestamp, s1 as symbol, b1 as bid, a1 as ask)
select  *,ffill(iif(prev(bid)==0,NULL,prev(bid))) as last_bid, ffill(iif(prev(bid)==0,NULL,prev(timestamp))) as last_bid_timestamp,  ffill(iif(prev(ask)==0,NULL,prev(ask))) as last_ask,  ffill(iif(prev(ask)==0,NULL,prev(timestamp))) as last_ask_timestamp from t