ERROR：Session/line number was not unique in database. History logging moved to new session.

 摘要：遇到此类错误，可以通过分段调试的方法找到引发错误的位置。

引发错误的原因不详，可能很基础。

---------------------------------------------------------------------------------------------------------------------------------

# Distribution of the peak number
file = pd.read_excel("......xls", sheet_name = "...")

DA_ratemap_allmice1 = getDaysAligned_Ratemap_allmice(maze_type = 1, isGood = 0)
DA_ratemap_allmice2 = getDaysAligned_Ratemap_allmice(maze_type = 2, isGood = 0)
print(np.shape(DA_ratemap_allmice1))
print(np.shape(DA_ratemap_allmice2))

# np.shape(DA_ratemap_allmice1) = (3955, 1296)
# np.shape(DA_ratemap_allmice1) = (2025, 1296)
# 自己生成一个试试，假定每个数组每一行都有一个最大值，最大值出现的位置为 x, x 属于[0,1296)
# 统计所有行中最大值在[0,1296)上出现的频数


def find_Decision_point(maze_type = 1):
    graph = {
        '1':[1,2],
        '2':[2,3,4],
        ...: [...],
        '144':[5,4,3],
    }
    co_path = [2,4,6,8,10]
    inc_path = [1,3,5,7,9]   #注意这是list，不是np.array()
    
    dp = []
    for key in graph.keys():
        # decision points
        if len(graph[key]) >= 3:
            dp.append(key)
            
    new_dp = np.zeros(len(dp))        
    # rank the decision point to new order:
    for i in range(len(dp)):
        if dp[i] in co_path:
            ind = np.where(co_path == dp[i])[0]
            new_dp[i] = ind+1
        else:
            ind = np.where(inc_path == dp[i])[0]
            new_dp[i] = ind + len(co_path)+1
    print(new_dp)
    return new_dp
print(find_Decision_point(1))

def CountDistributionOfPeak(DA_ratemap_allmice, day, DoyouwantPlot,maze_type):
    DA_ratemap_allmice_arr = np.array(DA_ratemap_allmice)
    DA_ratemap = DA_ratemap_allmice_arr[:,144*(day-1):144*day]
    peak_DA_ratemap = np.zeros_like(DA_ratemap)
    
    for i in range(len(DA_ratemap)):
        if np.sum(DA_ratemap[i])==0:
            continue
        peak_DA_ratemap[i][np.argmax(DA_ratemap[i])] = 1
    

    Dist = np.zeros(144)
    for i in range(np.shape(DA_ratemap)[1]):
        for j in range(np.shape(DA_ratemap)[0]):
            if peak_DA_ratemap[j][i]==1:
                Dist[i] += 1
    if DoyouwantPlot==0:
        return Dist
    else:
        dp = find_Decision_point(maze_type)
        plt.plot(range(1,145),Dist,'o',label = "other bins")
        plt.title("maze "+str(maze_type)+'  Day '+str(day))
        plt.xlabel('Spatial ID')
        plt.ylabel('Frequency')
        for i in range(len(dp)):
            plt.axvline(dp[i],color = 'black',linewidth = 1, ls = ':')
        plt.plot(dp,[Dist[dp[i]-1] for i in range(len(dp))], 'o', color = 'red',label = "decision points")
        plt.legend()
        return Dist

dist = CountDistributionOfPeak(DA_ratemap_allmice1,0,1,1)
print(dist)

maze_type = 1
fig = plt.figure(figsize=(48,24))
for i in range(9):
    print("Day",i+1)
    plt.subplot(2,5,i+1)
    dist = CountDistributionOfPeak(DA_ratemap_allmice1, i+1,1,maze_type)
plt.savefig(r"......",dpi=600)
plt.savefig(r"......",dpi=600)
plt.show()    

fig = plt.figure(figsize=(48,24))
maze_type = 2
for i in range(9):
    print("Day",i+1)
    plt.subplot(2,5,i+1)
    dist = CountDistributionOfPeak(DA_ratemap_allmice2, i+1,1,maze_type)
plt.savefig(r"......",dpi=600)
plt.savefig(r"......",dpi=600)
plt.show()

报错 

ERROR! Session/line number was not unique in database. History logging moved to new session

debug建议：分段debug，查找产生错误的位置。

# Distribution of the peak number
file = pd.read_excel("......xls", sheet_name = "...")

DA_ratemap_allmice1 = getDaysAligned_Ratemap_allmice(maze_type = 1, isGood = 0)
DA_ratemap_allmice2 = getDaysAligned_Ratemap_allmice(maze_type = 2, isGood = 0)
print(np.shape(DA_ratemap_allmice1))
print(np.shape(DA_ratemap_allmice2))

[out]
(3955, 1296)
(2025, 1296)

接着验证函数 find_Decision_point：

 bug出现了

说明该函数有问题。该函数由两部分组成，先验证其前一部分，没有问题。

再验证后一部分，发现有问题。（验证方法，用……注释掉其中一部分，然后再另一部分print()随便一些内容进行输出调试。）

之后发现，for循环内这句话有问题：

        print(np.where(co_path == dp[i])[0])
        if dp[i] in co_path:
            ind = np.where(co_path == dp[i])[0]   # co_path is list type
            new_dp[i] = ind+1                     # new_dp[i] = [] + 1

 原来，co_path是list类型，无法用np.where()找到对应值下标。ind的结果并非是一个int，而是[](empty list)。

只要前面一个cell出现了Session/line number...的error，后边都会出现。

奇怪的是，单独试验 [] + 1其实是会报错的，这是典型的TypeError

 所以，本文最终并没有解决这个bug如何出现，而是表明，这个bug有望通过修改代码本身即可解决，而不必像下文中所指的进行部分文件操作。当然这可能和不同具体情况相关，都是解决方案。ERROR! Session/line number was not unique in database. History logging moved to new session 178