ChatGPT给了三种方案

我用了第一种

   override fun onBindViewHolder(
        holder: RecyclerView.ViewHolder, position: Int, payloads: MutableList<Any>
    ) {

        when (getItemViewType(position)) {
            ITEM_VIEW_TYPE -> {
                val item = mModel.mListDatas[position] as ReadAfterBean
                if (holder !is ReadAfterSentenceHolder) {
                    return
                }
                holder.binding.apply {

                    for (payload in payloads) {
                        when (payload) {
                            in listOf("localReadsSentenceData") -> updateSentenceView(
                                item.localReadsSentenceData, this, position
                            )

                            in listOf("recordStatus") -> updateChangeStatus(
                                item.recordStatus, this, position
                            )

                            in listOf("localDialogActionData") -> updateActionData(
                                item.localDialogActionData, this, position
                            )

                            in listOf("playingData") -> updateSentencePlayingState(
                                item.playingData, this
                            )
                        }
                    }
                }


            }

            BOTTOM_STATUS_VIEW_TYPE -> {


            }

            else -> throw IllegalArgumentException("Invalid view type")
        }

        super.onBindViewHolder(holder, position, payloads)

    }

重写   holder: RecyclerView.ViewHolder, position: Int, payloads: MutableList<Any> 

三个参数的方案
payloads 这个 我从未用过。但是是真好用
 

这个方法重写后不要删除原来的 onBindViewHolder  还是照常设置数据，

当遇到频繁要刷新的数据时。

在需要频繁更新的数据中去设置。

        if (mModel.mListDatas[position] is ReadAfterBean) {
            val readAfterBean = mModel.mListDatas[position] as ReadAfterBean

            readAfterBean.recordStatus = data
            notifyItemChanged(position, "recordStatus")


        }

这样的数据不会频繁的干扰动画同UI的刷新。